Sensitive chiral analysis by CE: An update. Review

Laura Sánchez-Hernández Antonio Luis Crego María Luisa Marina Carmen García-Ruiz Departamento de Química Analítica, Facultad de Química, Universidad de Alcalá, Alcalá de Henares, Madrid, Spain Received July 20, 2007 Revised September 19, 2007 Accepted September 20, 2007 Review Sensitive chiral analysis by CE: An update A general view of the different strategies used in the last years to enhance the detection sensitivity in chiral analysis by CE is provided in this article. With this purpose and in order to update the previous review by García-Ruiz et al., the articles appeared on this subject from January 2005 to March 2007 are considered. Three were the main strategies employed to increase the detection sensitivity in chiral analysis by CE: (i) the use of off-line sample treatment techniques, (ii) the employment of in-capillary preconcentration techniques based on electrophoretic principles, and (iii) the use of alternative detection systems to the widely employed on-column UV–Vis absorption detection. Combinations of two or three of the above-mentioned strategies gave rise to adequate concentration detection limits up to 10210 M enabling enantiomer analysis in a variety of real samples including complex biological matrices. Keywords: CE / Chiral analysis / Detection sensitivity / In-capillary preconcentration DOI 10.1002/elps.200700531 Electrophoresis 2008, 29, 237–251 237 1 Introduction Due to the different biological activity that the enantiomers of a chiral compound may have, chiral analysis has nowadays an increasing interest in a variety of disciplines such as pharmaceutical, environmental, or food analysis, among others. The individual determination of the enantiomers of a chiral compound may require a high sensitivity for a great number of applications. Thus, the analysis of low concentrated and limited amounts of biological samples, the analysis of environmental samples where analytes are usually present at trace level or the determination of food components, ingredients, or residues are examples of this kind of applications. Moreover, the determination of enantiomeric impurities in drugs can require sensitive analytical methodologies due to the low percentages of the impurity that should be determined accordingtothe ICHguidelines (forimpurity contentshigher than 0.05% the impurities have to be reported) [1]. Since CE has shown in the last years a great potential to achieve chiral separations, the development of analytical methodologies enabling the sensitive determination of enantiomers has been the aim of a considerable number of articles. The use of at least one chiral selector in the separation media is necessary to enable a chiral separation by CE being CDs the favorite chiral selectors employed. Although there are some authors that consider as CZE the separation mode in CE when neutral CDs are used, we will consider in this review that when CDs are employed as chiral selectors, and independently of their nature, the separation mode is EKC. In fact, the enantiomeric discrimination in EKC is produced by a chromatographic mechanism where interactions are established between each one of the enantiomers and the chiral selector [2–4]. Although much less used, CEC and NACE can also be employed for chiral analysis. CEC is a hybrid technique between CE and HPLC characterized by a high separation efficiency due to the plug profile of the mobile phase driven by the EOF. A wide range of HPLC chiral stationary phases (CSPs) transferable to CEC or monolithic columns can be used to provide adequate enantioselectivity. Moreover, an important aspect from the point of view of the sensitivity and selectivity in chiral analysis is the easy coupling of CEC to MS detection, when the chiral selector in CEC is immobilized [5]. On the other hand, the Correspondence: Dr. Carmen García-Ruiz, Departamento de Química Analítica, Facultad de Química, Universidad de Alcalá, Ctra. Madrid-Barcelona Km. 33.600, E-28871 Alcalá de Henares, Madrid, Spain E-mail: [email protected] Fax: 134-91-8854971 Abbreviations: CBI, cyanobenz[f]isoindole; CSP, chiral stationary phase; DAS--CD, heptakis(2,6-diacethyl-6-sulfato)-b-CD; DIM, dimethindene; DIO, dioxopromethazine; ECL, electrochemiluminescence; FASS, field-amplified sample stacking; HS--CD, highly sulfated-b-CD; HS-ª-CD, highly sulfated-g-CD; HDAS-- CD, heptakis(2,3-di-O-acetyl-6-O-sulfo)-b-CD; LE, leading electrolyte; LLE, liquid–liquid extraction; MA, methamphetamine; MDA, methylenedioxyamphetamine; MDMA, methylenedioxymethamphetamine; MTD, methadone; NAC, N-acetyl L-cysteine; NE, norephedrine; OPA, orthophthalaldehyde; PHM, pheniramine; poly-L-SUCL, poly(sodium N-undecenoxy carbonyl-L-leucinate); poly-LL-SUCLV, poly (sodium N-undecenoxycarbonyl-LLleucyl-valinate); SPCD, sample preconcentration with chemical derivatization; TEA, triethylamine; TE, terminating electrolyte; TMD, tramadol ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 238 L. Sánchez-Hernández et al. Electrophoresis 2008, 29, 237–251 aqueous buffer employed in CE is replaced in NACE by an organic solvent containing an electrolyte. The use of a solvent instead of an aqueous buffer has provided additional selectivities to those obtained in aqueous CE systems and may be advantageous for the hyphenation of this separation mode with MS, if volatile solvents are used [6]. A variety of strategies have been employed in CE in order to obtain the detection sensitivity needed for a given application. These strategies included the use of off-line or on-line sample treatment techniques, sample preconcentration in the capillary using techniques based on electrophoretic principles, and/or the use of alternative detection systems to the UV–Vis absorption detection [7]. In addition, other strategies such as partial filling of the capillary or the use of CDs with countercurrent migration can be necessary, if MS detection is employed when chiral selectors are in the separation media. These strategies avoid the introduction of these compounds into the MS detector which causes a damage in electrospray efficiency and increases background noise decreasing the sensitivity of detection. Some reviews have been published in the last years covering the chiral separation of drugs [8–10] or pollutants and their metabolites [11]. The aim of this review is to provide a general view of the different strategies that have been used in the last years in order to enhance the sensitivity of detection in chiral analysis by CE with application to the analysis of drugs or biological, environmental, or food samples. Articles appeared from the publication of the previous review by García-Ruiz and Marina [7] have been considered covering the period of time from January 2005 to March 2007. 2 Enhancement of the sensitivity in chiral analysis by CE An important aspect to take into account in CE is the very low volumes injected in the system (in the nL range) that imply that the detection of a concentration 1025M leads to the detection of ,10214 mol of analyte. For this reason, all those works where at least one strategy has been employed to improve the detection sensitivity in chiral analysis by CE enabling the determination of enantiomers at least at molar concentrations of 1025have been included in this review. Next sections will describe the different approaches used in the last years to enhance the detection sensitivity in chiral analysis by CE: (i) off-line sample treatment techniques, (ii) in-capillary preconcentration techniques, and (iii) alternative detection systems to the widely employed on-column UV– Vis absorption detection. 2.1 Off-line sample treatment techniques in sensitive chiral analysis by CE Table 1 groups the different sample treatments employed in the period of time reviewed in this article prior to chiral analysis by CE. Analytes, samples, separation buffer, detection system, and detection limits (LODs) obtained are also given in this table. Sample treatments can be aimed to eliminate some components of the sample matrix in addition to enhance the detection sensitivity through sample preconcentration. SPE, liquid–liquid extraction (LLE), solid-phase microextraction (SPME), or microdialysis were used. In all cases, off-line strategies were used. SPE and LLE were the main extractive techniques employed. Preconcentration by both techniques implies the reconstitution of the residue obtained after sample treatment in a volume smaller than the initial sample volume. Therefore, the improvement in the concentration sensitivity will depend on the sample volume available. However, preconcentration possibilities by SPE are usually better than by LLE, because the volume factor (sample volume/residue volume) is more favorable in SPE than in LLE. SPE is one of the most popular and widely used extractive techniques used for liquid samples due to its high selectivity and also preconcentration possibilities. SPE was only employed as sample treatment to isolate chiral drugs in biological samples as urine [12, 13, 15, 17] and to selectively preconcentrate chiral herbicides in spiked water samples [14, 16]. Most of these works were performed with UV detection [12–14, 17] achieving LODs in the 1028–1027M range. These LODs were improved by one or two orders of magnitude using SPE as sample treatment depending on the sample volume available (preconcentration factors ranging from 3 to 250 were achieved). However, other detection systems such as ESI-MS were also used enabling to detect up to 361028and 661028M for each salbutamol enantiomer [15]. In this case, the SPE step enabled an enrichment of the initial sample concentration of four times. The best LODs (261029M) were reached using fluorescence detection for the analysis of glufosinate enantiomers [16]. The high sensitivity achieved in the latter work was due to the combination of an SPE step with an in-capillary preconcentration strategy. Classical LLE was also employed for the extraction and preconcentration of chiral compounds prior to CE chiral analysis. Thus, Table 1 shows the use of LLE for the determination of drugs in biological samples as plasma [20–23], human serum [17, 19], human urine [17], and a microsomal fraction of liver homogenates [18]. Preconcentration factors up to 21 were achieved by this extraction procedure. After this sample treatment step, LODs from 1026to 361028M were reported when UV detection was used [17–20] and LODs ranging form 361026to 461029M were reached with ESI-MS detection [21, 22]. As example, Schappler et al. [22] assessed two approaches to enhance the sensitivity in CE. The former consisted of a protein precipitation using ACN followed by hydrodynamic injection of the supernatant. The second was the combination of LLE, which produces a sample cleanup and enrichment minimizing any matrix effect, with electrokinetic injection. In spite of the fact that the former method was rapidly achieved with minimal ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2008, 29, 237–251 CE and CEC 239 Table 1. Off-line sample treatment techniques employed for the enhancement of the sensitivity in chiral analysis by CE Sample treatment Analyte and sample Separation buffer Detection LOD (M) Ref. SPE Lorazepam in human urine 6 mM borate/10 mM phosphate (pH 9.1) 160 mM HP-b-CD 175 mM SDS UV-200 nm ,261025[12] SPE CIT, DCIT, DDCIT, CIT-NO, and CIT-PA in human urine 20 mM phosphate (pH 5) 10.2% CM-g-CD 10.05% HPMC UV-205 nm 661028– 361027 [13] SPE Malathion in spiked tap water 25 mM Tris (pH 7.0) 120 mM CM-b-CD UV-230 nm 661027[14] SPE Salbutamol in human urine 10 mM ammonium formate (acidified with 0.75 M formic acid) 115 mM HDAS-b-CD MS 361028, 661028 [15] SPE DNS-DL-Glufosinate in spiked river water 2 mM phosphate (pH 6.5) 117 mM g-CD Fluorescence (lexc = 327 nm, lem = 557 nm) 261029[16] SPE/LLE Ibuprofen in human serum and urine 200 mM orthophosphoric acid 1200 mM triethanolamine (pH 5.0) 150 mM TM-b-CD UV-220 nm ,261027 (serum), ,1026 (urine) [17] LLE Hydroxychloroquine and its metabolites (DCQ, DHCQ, BDCQ) in microsomal fraction of liver homogenates 100 mM Tris/phosphate (pH 9.0) 11% HS-b-CD 130 mg/mL HP-b-CD UV-220 nm ,1027[18] LLE Propafenone in human serum 100 mM phosphate (pH 2.0) 10.6% HS-b-CD. UV-195 nm ,361028[19] LLE Ketamine and norketamine in equine plasma 50 mM Tris (pH 2.5) 110 mg/mL HS-b-CD UV-195 nm ,461028[20] LLE Amphetamine derivatives (A, MA, MDA, MDMA, MDEA, E, NE) in plasma 20 mM ammonium formate (pH 2.5) 10.15% HS-g-CD MS ,761027– 361026 [21] LLE Amphetamine derivates (A, MA, MDA, MDMA, MDEA, TMD, MTD) in human plasma 20 mM ammonium formate (pH 2.5) 10.15% HS-g-CD MS 461029[22] LLE Disopyramide in spiked plasma sample 40 mM acetate (pH 4.5) 13 mg/mL HS-b-CD ECL ,1027[23] SPME (1R,2S)-Ephedrine, (1R,2R)- pseudoephedrine, (1S,2S)- pseudoephedrine in water and human urine 150 mM phosphate (pH 2.5) 117.5 mM b-CD UV-192 nm 2–361028[24] A, amphetamine; BDCQ, bisdesethylchloroquine; CIT, citalopram; CIT-NO, citalopram N-oxide; CIT-PA, citalopram propionic acid; CM-g-CD, carboxymethyl-g-CD; DCIT, demethylcitalopram; DCQ, desethylchloroquine; DDCIT, didemethylcitalopram; DHCQ, desethylhydroxychloroquine; DNS, dansyl choride; E, ephedrine; HDAS-b-CD, heptakis(2,3-di-O-acetyl-6-O-sulfo)-b-CD; HP-b-CD, 2-hydroxypropyl-b-CD; HPMC, hydroxypropylmethylcellulose; MDEA, methylenedioxyethylamphetamine; NDA, naphthalene-2,3-dicarboxaldehyde; TM-b-CD, heptakis 2,3,6-tri-O-methyl-b-CD. sample manipulation, a concentration factor of 1000-fold was obtained with the second one although it was more time consuming. Finally, the MS signal suppression effect was investigated on the complex matrices of biological samples (plasma) with a conventional CE-ESI-MS setup. Suppression occurred in the migration window of analytes of interest in conventional analysis. However, when LLE was employed, no matrix effects were evidenced leading to the conclusion that this sample preparation method remains of utmost interest for the analysis of biological samples by CE-MS. The combination of LLE-electrokinetic injection in CE-MS enabled to obtain the best LODs (,1029M) indicated in Table 1, and this was for amphetamine derivatives in human plasma. Figure 1 shows the sensitive determination of seven am- ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 240 L. Sánchez-Hernández et al. Electrophoresis 2008, 29, 237–251 Figure1. Chiralanalysisofaplasmasamplespikedwith0.5 ppbof each enantiomer of seven different amphetamine derivatives, after LLE and electrokinetic injection in CE-ESI-MS. Electrophoretic conditions: fused-silica capillary, 75 cm total length and 50 mm id; separation buffer, 20 mM ammonium formate (pH 2.5) containing 0.15% HS-g-CD; separation temperature, 257C; applied voltage, 25 kV; electrokinetic injection, 10 kV610 s. MS conditions: sheath liquid, isopropanol/H2O (50:50 v/v) containing 0.5% formic acid; syringe flow rate, 3 mL/min. ESI capillary at 4.5 kV. The nebulizing pressure and the drying gas flow rate were set at 4 psi and 4 L/min, respectively. Gas temperature, 2007C; fragmentor 70 V. Identification of peaks: A, amphetamine; MA, methamphetamine; MDA, methylenedioxyamphetamine; MDMA, methylenedioxymethamphetamine; MDEA, methylenedioxyethylamphetamine; TMD, tramadol; MTD, methadone (reprinted from ref. [22] with permission). phetamine derivatives in a plasma sample spiked with 0.5 ppb of each enantiomer after LLE and electrokinetic injection in CE-ESI-MS. On the other hand, Fang et al. [24] proposed an SPME sample treatment previous to the injection into a CE system with UV detection as a chiral sensitive method for doping control of ephedrine derivatives. Although this extractive technique is mainly used in GC where compounds are desorbed from the fiber by applying high temperatures, in this work the analytes (ephedrine derivatives) were extracted from 5 mL of the liquid samples. This was carried out by exposing the coated fiber end to the headspace of the vial for an appropriate time and then desorbing the analytes in 80 mL of back-extraction solvent containing an optimized concentration of ACN. Figure 2 shows a ,160-fold intensity improvement after SPME of ephedrine derivatives from an urine sample. Interestingly, SPME not only enhanced the sensitivity of detection by concentrating the sample, also supplied the possibility to employ an in-capillary sample preconcentration strategy providing an appropriate sample matrix to be injected in the CE system. Figure 2. Separation of ephedrine derivatives in a spiked urine sample (a) directly injected in CE (concentration of each analyte: 5.00 mg/mL), and (b) after off-line SPME (concentration of each analyte: 0.25 mg/mL) followed by CE. Electrophoretic conditions: fused-silica capillary, 60.5 cm total length and 75 mm id; separation buffer, 150 mM phosphate buffer (pH 2.5) containing 17.5 mM b-CD; separation temperature, 207C; applied voltage, 25 kV; electrokinetic injection, 7 kV610 s. UV detection at 192 nm. Identification of peaks: (2)-PE, (1R,2R)-pseudoephedrine; (2)-E, (1R,2S)-ephedrine; (1)-PE, (1S,2S)-ephedrine (reprinted from [24] with permission). 2.2 In-capillary preconcentration techniques based on electrophoretic principles in chiral analysis by CE Another way to increase the sensitivity in chiral CE is by using in-capillary preconcentration techniques based on electrophoretic principles. Table 2 summarizes the main characteristics of works where two different in-capillary preconcentration strategies were used. On one hand, classical in-capillary preconcentration techniques, such as ITP, stacking, and/or sweeping were employed, and on the other hand, in-capillary sample preconcentration with chemical derivatization (SPCD) is included in this table as an innovative strategy. A brief description of the different preconcentration techniques will be made previously to indicate the applications performed by CE in the chiral field. In ITP, ionic analytes are concentrated and separated on the basis of their electrophoretic mobilities using a discontinuous buffer system (leading electrolyte (LE) and tailing electrolyte (TE). This technique may be performed in the same capillary where the electrophoretic separation is ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2008, 29, 237–251 CE and CEC 241 Table 2. In-capillary preconcentration techniques employed for the enhancement of the sensitivity in chiral analysis by CE Preconcentration technique Analyte and sample Separation buffer Detection LOD (M) Ref. cITP Antihistaminic drugs (PHM, DIM, DIO) in urine 25 mM acetic acid (pH 3.2–4.5) 12.5 mg/mL (PHM, DIM) and 5.5 mg/mL (DIO) CE-b-CD UV-265 and 240 nm 461029– 261028, 1028 [25] cITP Drugs standards (S-alprenolol, S-atenolol, R-propranolol, R-salbutamol, S-terbutaline) LE: 160 mM acetate (pD 4.7) 110 mM b-CD TE: 160 mM acetic acid (pH 2.4) 110 mM b-CD NMR-600 MHz ,2.561024[26] tITP R,S-Timolol in standard 40 mM KOH in methanol/ethanol (40:60 v/v) 1100 mM (1)-KPA UV-220 461026[27] FASS 561026 FASS (1R,2S)-Ephedrine, (1R,2R)- pseudoephedrine, (1S,2S)- pseudoephedrine in water and human urine 150 mM phosphate (pH 2.5) 117.5 mM b-CD UV-192 nm 2–361028[24] LVSS DNS-DL-glufosinate in spiked river water 2.0 mM phosphate (pH 6.5) 117 mM g-CD Fluorescence (lexc = 327 nm, lem = 557 nm) 261029[16] LVSS 1 sweeping CBI-Ser, CBI-Glu in biological samples 25 mM phosphate (pH 2.0) 12% HS-b-CD LIF (lexc = 420 nm) 2610210, 3610210 [28] Sweeping Lorazepam in human urine 6 mM borate/10 mM phosphate (pH 9.1) 160 mM HP-b-CD 175 mM SDS UV-200 nm ,261025[12] In-capillary SPCD OPA/NAC-amino acids (Ala, Glu) in E. coli bacterial culture 140 mM borate (pH 9.5) 11mMb-CD UV-340 nm 461027, 661027 [29] In-capillary SPCD OPA/NAC-muramic and diaminopimelic acids in E. coli bacterial culture 140 mM borate (pH 9.5) UV-214 and 340 nm 261026, 261027 [30] CE-b-CD, carboxyethyl-b-CD; DIKGA, di-O-isopropylidene-2-keto-L-gulonic acid; DIM, dimethindene; DIO, dioxopromethazine; KPA, ketopinic acid; PHM, pheniramine. achieved, which is denominated transient ITP (tITP), or in two different capillaries, which is named capillary ITP (cITP). Table 2 shows that cITP was used as preconcentration technique to determine different antihistaminic drugs in a biological sample as urine [25] and to detect several drugs in standard solutions by CE-NMR [26]. Likewise, this table also shows the use of tITP for the preconcentration of R,S-timolol [27]. Thus, Mikus et al. [25] proposed a highly sensitive cITPEKC method combining two coupled capillaries: in the first capillary, ITP was carried out to achieve a preseparation and a sample preconcentration, and in the second capillary, where the sample was transferred on-line, the EKC separation occurred. These experiments showed favorable conditions for the separation and determination of traces (ng/mL) of antihistaminic drugs (dioxopromethazine (DIO), dimethindene (DIM), and pheniramine (PHM)) enantiomers present in urine samples diluted with water. By using UV detection, LODs achieved in this work were from 461029to 261028M depending on the compound analyzed. Another example of the use of cITP sample stacking process is the separation and concentration of analytes for NMR measurements [26]. It is important to emphasize that the coupling of cITP to NMR enabled to detect small amounts (,2 nmol) of the basic drugs studied, but the concentrations detected (,70 mM) were far away of a sensitive detection. Finally, a rapid analytical method by using tITP for the determination of the enantiomeric impurity of S-timolol was developed by Hedeland et al. [27]. In this method, the combination of 1S,4R-(1)- ketopinic acid (KPA) as chiral selector in NACE and preconcentration by tITP provided LODs as low as 0.2% of Rtimolol in S-timolol samples. With respect to stacking preconcentration, it is produced because ions migrate electrophoretically through a low-conductivity sample matrix into a high-conductivity buffer solution and they are focused in a thick zone between the boundaries of both solutions. Most stacking modes use ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 242 L. Sánchez-Hernández et al. Electrophoresis 2008, 29, 237–251 hydrodynamic injection of large-volume sample being named large-volume sample stacking (LVSS). However, when electrokinetic injection is used to focus the analytes of the sample, the strategy is denominated field-amplified sample stacking (FASS). As shown in Table 2, highly sensitive chiral methods were developed using this preconcentration technique based on the stacking of the analytes in narrow bands previously to the electrophoretic separation. The possibility of preconcentration by FASS was evaluated and compared with tITP by Hedeland et al. [27]. FASS was performed using an electrokinetic injection of 8 kV for 10 s. Similar LODs, in the 1026M range, were obtained by both in-capillary preconcentration techniques. In addition, FASS was applied for in-capillary sample concentration of ephedrine derivatives in water and urine samples [24]. In this work, a study of the performance of the FASS procedure (7 kV for 10 s) using the samples prepared in water enabled to obtain a concentration factor of up to 80-fold compared with the conventional electrokinetic sample injection (7 kV for 10 s), where the samples were dissolved in the separation buffer. On the other hand, the concentration enrichment achieved by LVSS allowed the analysis of traces of the enantiomers of the herbicide glufosinate in spiked river water (model sample) [16]. The combination of this in-capillary preconcentration strategy with SPE enabled to obtain LODs for Dand L-glufosinate enantiomers as low as 261029M. Moreover, Kirschner et al. [28] obtained a very high sensitivity with a CE-LIF method involving a combination of LVSS and sweeping which is a preconcentration technique enabling a significant increase in the detection sensitivity for those analytes with a high solute-pseudostationary phase association constant [31]. The combined preconcentration mechanism proposed by Kirschner et al. is illustrated in Fig. 3A. First, LVSS involves a combination of field-amplified stacking and pH-mediated stacking. Then, sweeping using the anionic highly sulfated-b-CD (HS-b-CD) as pseudophase interacting with the analytes was performed. A solution containing the analytes was injected filling 1/3 of the capillary. Then, reverse polarity was applied and the anionic cyanobenz[f]isoindole (CBI)–amino acids migrated toward the pH junction at the outlet side of the injection plug, where they were substantially neutralized and stacked by the low pH buffer. The migration of the analyte anions produced a field-amplified stacking in the dilute water. Simultaneously, EOF began to pump water out of the capillary and moved the stacked analytes band toward the inlet. During this period, most electrolytes stayed within the capillary, since the current was near zero or a few microamperes. Finally, once the analytes were nearly ejected from the capillary, the current increased as the HS-b-CD migrated rapidly through the stacked band of the analyte, sweeping it to the outlet. Interestingly, LODs up to 10210 M for the baseline resolved CBI– amino acid enantiomers were reached by this way. This preconcentration technique was also applied to more complex mixtures of amino acids without loss of resolution, as it can be seen in Fig. 3B. Another example of sweeping is the senFigure 3. (A) Schematic diagram of the stacking/sweeping: (i) hydrodynamic injection of large volume (1/3 of the capillary) of the CBI–amino acids in water at pH 6.0; (ii) migration toward the pH junction at the outlet side of the injection plug of the anionic CBI–amino acids; (iii) pumping water out of the capillary and movement of the stacked analyte band toward the inlet by the EOF; (iv) sweeping of the HS-b-CD through the stacked band of analyte. (B) Electropherogram showing the potential of stacking/ sweeping-EKC combination to enantioseparation of a complex sample of CBI–amino acids (,0.5 mM each). Electrophoretic conditions: fused-silica capillary, 70 cm total length (45 cm detector length) and 25 mm id; separation buffer, 25 mM phosphate buffer (pH 2.0) containing 2% HS-b-CD; applied voltage, 230 kV; hydrodynamic injection, 380 mbar for 180 s. LIF detection with lexc at 420 nm. Peak identification: 1, CBI-D-arginine (Arg); 2, CBI-L-Arg; 3, CBI-D-histidine (His); 4, CBI-L-His; 5, CBI-glycine (Gly); 6, CBI-Ltyrosine (Tyr); 7, CBI-L-glutamine (Glu); 8, CBI-D-Ser; 9, CBI-L-Ser; 10, CBI-L-Glu; 11, CBI-D-Glu (adapted from ref. [28] with permission). sitive determination of lorazepam in human urine, a drug used for the treatment of anxiety and with sedative and hypnotic properties. In this work, the sweeping of the sample was achieved with a buffer containing SDS micelles and a ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2008, 29, 237–251 CE and CEC 243 neutral CD as chiral selector [12]. Sweeping was performed by preparing the sample in a matrix without micelles but otherwise similar to the running buffer, and prolonging the sample zone by applying longer injection times. In this way, when voltage was applied, micelles sweep the analyte in a narrow band producing a sample preconcentration. Table 2 also includes two works where a single-step method that combines in-capillary SPCD was developed for the enantioselective determination of amino acids [29] and the bacterial biobacters muramic acid and diaminopimelic acid [30] in Escherichia coli bacterial culture using CE with UV detection. The effective separation of eight amino acid enantiomers using in-capillary derivatization with orthophthalaldehyde (OPA)/N-acetyl L-cysteine (NAC) was performed [29]. In comparison with conventional CE, SPCD provided a 40fold improvement in concentration sensitivity and permitted shorter total analysis times maintaining chiral resolution due to the lower average mobility of zwitterionic amino acids in the weakly acidic sample plug. Figure 4A shows the general principle of SPCD in CE for the separation of the enantiomers of the two amino acids, alanine (Ala) and glutamic acid (Glu). Briefly, the injection sequence was carried out by first rinsing the capillary with borate buffer that served as an optimal alkaline medium for analytes and OPA and NAC as derivatizing agents. A multiple sample hydrodynamic injection sequence was performed consisting of short concentrated plugs of NAC and OPA reagents positioned in between a long plug of dilute sample. OPA was injected at the back end of the sample plug because it is neutral and comigrates with the EOF, unlike the anionic chiral NAC coreagent that migrates with a slower apparent mobility. The sample was injected with low pressure dissolved in phosphate buffer to reduce the local electrophoretic mobility of weakly ionic metabolites in the sample to induce electrokinetic focusing. Consecutively, OPA and NAC zones comigrated and preconcentration sample zones resulted in the formation of diastereomeric isoindole adducts. Afterward, enantiomeric resolution of the adducts was achieved along with increasing band separation of all species. Although the separation of the diastereomers of the amino acids was possible without chiral additives (Fig. 4B), the addition of 1 mM b-CD allowed the resolution of the last amino acid without deterioring sample enrichment properties, but decreasing resolution of the two first migrating amino acids (Fig. 4C). The SPCD method developed integrated sample enrichment with chemical labeling steps directly within a single capillary during electromigration improving the concentration sensitivity but retaining high-resolution chiral separations. LODs of 261027M for muramic acid, and 261026M for diaminopimelic acid, were achieved by this method. Nevertheless, in this case, the use of increasing concentrations of b-CD as a neutral chiral selector to the run buffer was unsuccessful for further improving of diaminopimelic acid resolution [30]. Figure 4. (A) General principle of SPCD–CE for single-step enantioselective analysis of submicromolar levels of amino acids: (i) multiple hydrodynamic injection sequence; (ii) on-line sample preconcentration; (iii) in-capillary chemical labeling by zone passing of OPA/NAC; (iv) chiral separation of diastereomeric amino acid adducts formed. (B) Electropherogram of eight amino acid enantiomers using in-capillary OPA/NAC derivatization of amino acid adducts by SPCD–CE. (C) Electropherogram demonstrating the use of 1 mM b-CD as a chiral additive to enhance the enantioselectivity by dynamic inclusion complexation of amino acid adducts by SPCD–CE. Electrophoretic conditions: fused-silica capillary, 65 cm total length and 50 mm id; separation buffer, 140 mM borate buffer (pH 9.5) containing 1 mM b-CD for Fig. (C); separation temperature, 257C; applied voltage, 25 kV; hydrodynamic injection, 35 mbar6100 s; UV detection at 340 nm. Sample solutions contained 25 and 50 mMoftheD-andL-amino acids, respectively. Analyte peak numbers correspond to amino acid–isoindole adducts: 1a, D-Ser; 1b, L-Ser; 2a, D-Ala; 2b, L-Ala; 3a, D-Glu; 3b, L-Glu; 4a, D-Asp; 4b, L-Asp; *, OPA hydrolysis products (adapted from ref. [29] with permission). ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 244 L. Sánchez-Hernández et al. Electrophoresis 2008, 29, 237–251 2.3 Alternative detection systems to on-column UV–Vis absorption detection for the enhancement of the sensitivity in chiral analysis by CE UV–Vis absorbance detection is the first option to be considered in CE given its interesting features such as commercial availability, simplicity, versatility, relatively low cost, and frequent use as universal detection technique because many organics can be detected at 195–210 nm. This detector produces LODs in CE corresponding to a few femtomoles of analyte (at subpicomole levels), i.e., high mass sensitivity. However, due to the need of small volumes employed in CE to avoid peak broadening that decrease the efficiency of the separation (pL to nL volumes), such sensitivity is in the micromolar range, i.e., appears modest when expressed in terms of concentration (LODs ranging from 1025to 1027M depending upon the analyte being analyzed). These LODs are clearly insufficient to solve many analytical problems. One option to overcome the poor concentration sensitivity obtained in CE is the selection of low-UV wavelengths (190– 205 nm), where this detection exhibits the best sensitivity, possibility used in several works summarized in Table 1. Another option to improve detection sensitivity in CE is the use of alternative detection systems. During the period of time reviewed in this article (see Table 3), LIF and MS detection systems were mainly employed. LIF system is configured to make the detection directly on the separation capillary (on-column detection), such as UV–Vis absorption detection, but the second system is connected to the end of the capillary (end-column detection). In addition, electrochemiluminescence (ECL) was also employed. 2.3.1 On-column detection in chiral analysis by CE LIF detection is one of the most sensitive on-column detection systems currently available in which it is possible to use several lasers, the common argon laser (458 and 488 nm), the He-Cd laser (440 nm), or the blue diode laser (420 nm). Since only a few chiral compounds possess native fluorescence, derivatization procedures are usually required to detect other nonfluorescent analytes. LIF was used for the chiral determination of photosensitizers [32], drugs in human plasma [33], and anticancer agents in drugs [34] achieving LODs around 1028M. However, the main application of this detection system in CE was the analysis of amino acid enantiomers in standards [35], foods [36], and biological samples [28, 37–42]. In all cases, the methods involved precapillary chiral derivatization of the amino acids with different derivatizing reagents, FITC [35, 36], 4-fluoro-7-nitro2,1,3-benzoxadiazole (NBD-F) [37], CBI [28, 38–41], or fluorescamin (FA) [42]. Chiral and sensitive analysis of amino acids is a remarkable methodology that can provide important information on adulteration and quality of food products [49]. For example, analysis of chiral amino acids in vinegars was shown to be a powerful method to detect different adulterations including the detection of synthetic vinegars [50]. Thus, several D-amino acids were detected and quantified in vinegars by MEKC-LIF, observing interesting differences in their Land D-amino acid profiles and contents [36]. LODs lower than ,20 nM were achieved using FITC as derivatizing reagent. This fluorescent label has been frequently chosen since its excitation wavelength matches the 488 nm light of the argon laser, and the derivatives are easily formed and generate strong fluorescence signals [51]. Several improved methods were developed for the sensitive determination of D-serine (Ser) in neural samples [28, 37–39]. This molecule is a primary endogenous amino acid that binds to the glycine site of N-methyl-D-aspartate (NMDA) receptor involved in a variety of physiological functions and disorders including memory, learning, pain, and ischemia. The sensitive detection of this amino acid is essential, because it is at low concentration in biological samples. Thus, two different EKC methods with LIF detection for the chiral separation of D/L-Ser were developed and applied to detect D-Ser in higher vertebrates, such as rats, achieving LODs about 1027M [37, 38]. In one of the methods, a saccharide [D-(1)-glucose] was added to the separation buffer in order to enhance the chiral recognition of 2-hydroxypropyl-g-CD (HP-g-CD) [38]. On the other hand, a procedure for the determination of D-Ser in squirrel brain was proposed to achieve a very high sensitivity (LODs up to 10210 M) [28]. In this case, the process involved two preconcentration techniques (stacking and sweeping), as has previously been commented. By this method, in addition to L-Ser, appreciable levels of D-Ser, L-aspartate (Asp), and L-glutamate (Glu) were observed in microdialysate from the hippocampus of artic ground squirrels, while D-Asp and D-Glu were below the LOD. The enantioseparations were accomplished with HS-b-CD as chiral selector at low pH and reverse polarity. D-Ser was also detected in invertebrates [38, 39]. Thus, Zhao et al. [39] studied the contents of D-Ser in Aplysia californica, a sea mollusc widely used as neuronal model. The separation of CBI-D/L-Ser enantiomer was achieved by using a dual chiral selector system consisting of b-CD and chiral micelles formed by DOC. For the first time, peaks corresponding to L-Ser and D-Ser were well identified in Aplysia ganglian homogenates. It was noticed that while the levels of L-Ser were similar, D-Ser levels varied substantially from animal to animal. Interestingly, D-Ser was not detected in single neurons isolated from Aplysia ganglia due to D-Ser might perhaps not occur in neurons of Aplysia or the contents in single neurons were too low to be detected. This lack of sensitivity was reported also in other work of the same authors [38]. D-Asp was also detected in the central nervous system of A. californica [40, 41]. There are some indications that D-Asp is a neuromodulator, hormone, or even a precursor for the endogenous synthesis of the NMDA receptor. Miao et al. [40] demonstrated an approach for the quantitative investigation ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2008, 29, 237–251 CE and CEC 245 Table 3. Alternative detection systems to on-column UV–Vis absorption detection employed for sensitive chiral analysis by CE Detection System CE Mode Analyte and sample Separation buffer LOD (M) Ref. LIF (lexc = 422/488 nm, lem = 690 nm) CD-MEKC Porphyrin and phthalocyanines in standards 200 mM borate (pH 9.2) 1 10 mM HP-b-CD 110 mM SDS ,361028[32] LIF (lexc = 442 nm, lem = 500 nm) CD-EKC CBI-Baclofen in human plasma 50 mM borate (pH 9.5) 12% HS-b-CD 561028[33] LIF (lexc = 320 nm, lem = 380–600 nm) CD-EKC Homocamptothecin derivatives in drugs 75/25 mM phosphate (pH 2.5) 17.5% w/v HS-b-CD/2.5% w/v HS-b-CD ,1028[34] LIF (lexc = 488 nm, lem = 520 nm) EKC FITC-amino acids (Glu, Pro) in standards 50 mM acetate (pH 6.1) 1 1.25 mM vancomycin ,1025[35] LIF (lexc = 488 nm, lem = 520 nm) CD-MEKC FITC-amino acids (Arg, Pro, GABA, Ala, Glu, Asp) in vinegars 100 mM borate (pH 9.7) 1 20 mM b-CD 130 mM SDS ,261028[36] LIF (lexc = 457.9 nm) CD-EKC NBD-F-D-Ser in rat brain 100 mM borate (pH 10.0) 1 40 mM HP-b-CD 361027[37] LIF (lexc = 457.9 nm) CD-EKC CBI-D-Ser in rat brain and mollusc neurons 60 mM borate (pH 10.0) 12M urea 115% D-glucose 1 20 mM HP-g-CD 1027[38] LIF (lexc = 420 nm) CD-EKC CBI-Ser, CBI-Glu in hippocampus of squirrel 25 mM phosphate (pH 2.0) 12% HS-b-CD ,10210 [28] LIF (lexc = 457.9 nm) CD-MEKC CBI-D-Ser in mollusc neurons 100 mM borate (pH 9.5) 1 30 mM b-CD 160 mM DOC 361028[39] LIF (lexc = 457.9 nm) CD-MEKC CBI-D-Asp in mollusc neurons 50 mM borate (pH 9.4) 1 20 mM b-CD 150 mM SDS 5610210 [40] LIF (lexc = 457.9 nm) CD-MEKC CBI-D-Asp in mollusc neurons 50 mM borate (pH 9.4) 1 20 mM b-CD 150 mM SDS ,561026[41] LIF (lexc = 457.9/488 nm) CD-MEKC FA-D-amino acid-containing neuropeptides in mollusc neurons 50 mM borate (pH 9.4) 1 20 mM g-CD 150 mM SDS 861028[42] Fluorescence (lexc = 327 nm, lem = 557 nm) CD-EKC DNS-DL-Glufosinate in spiked river water 2.0 mM phosphate (pH 6.5) 117 mM g-CD 261029[16] MS CD-EKC Amphetamine derivatives (A, MA, MDA, MDMA, MDEA, E, NE) in plasma 20 mM ammonium formate (pH 2.5) 10.15% HS-g-CD ,761027– 361026 [21] MS CD-EKC Amphetamine derivates (A, MA, MDA, MDMA, MDEA, TMD, MTD) in plasma 20 mM ammonium formate (pH 2.5) 10.15% HS-g-CD 461029[22] MS CD-EKC MA, AP, DMA, E, NE, ME in human urine 1 M formic acid (pH 1.7) 10.85 mM DAS-b-CD ,661028– 1027 [43] MS MEKC b-Blockers (Ate, Met, Pin, Oxp, Alp, Pro, Car, Tal) in standards 25 mM ammonium acetate 125 mM TEA (pH 8.0) 115 mM poly-L-SUCL 961027– 761026 [44] MS MEKC Lorazepam, oxazepam, and nefopam in standards 25 mM ammonium acetate (pH 8.0) 115% ACN 115 mM poly-L-SUL ,661026[45] Lorazepam and oxazepam in standards 25 mM ammonium acetate (pH 8.5) 115 mM poly-LLSUCLV ©2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com