ANALYTICAL SCIENCES FEBRUARY 2005, VOL. 21 115 2005 © The Japan Society for Analytical Chemistry

Chiral CE Separation of -Derived

Zhe QUAN,* Yaru SONG,* Gladys PETERS,* Ming SHENWU,* Yinghong SHENG,* Huey-Min HWANG,** and Yi-Ming LIU*†

*Department of Chemistry, Jackson State University, 1400 Lynch St., Jackson, MS 39217, USA **Department of Biology, Jackson State University, 1400 Lynch St., Jackson, MS 39217, USA

An enantiomeric separation of dopamine-derived neurotoxins by capillary electrophoresis has been developed. (TIQ), dopamine (DA), (R/S)-1-benzyl-TIQ (BTIQ), (R/S)-6,7-dihydroxy-1-methyl-TIQ (salsolinol, Sal), and (R/S)-6,7-dihydroxy-1, 2-dimethyl-TIQ (N-methyl-salsolinol, NMSal) were studied as model compounds. The CE running buffer (50 mM phosphate buffer at pH 3.0) contained 1.5 M urea and 12 mM β-CD as a chiral selector. During separation, the (R)-enantiomers formed more stable inclusion complexes with β-CD, and thus had a longer migration time than their optical antipodes. It was noticed that the recovery rates of these TIQ derivatives were very poor (< 15%) during protein precipitation, a procedure widely used for cleaning up biological samples. The recovery was significantly improved by pre-mixing the sample with a surfactant (e.g., sodium hexanesulfonate or Triton X-100) to reduce the co-precipitation. The present method in combination with electrospray ionization tandem mass spectrometry (ESI-MS/MS) was applied to study samples obtained from in vitro incubation of two catecholamines, dopamine and epinine, with aldehydes forming neurotoxins including (S)- and (R)-NMSal enantiomers. The later is known to induce Parkinsonism in rats.

(Received April 6, 2004; Accepted June 7, 2004)

neuroblastoma SH-SY5Y cells.18 It was shown that endogenous Introduction TIQ formation and TIQ N-methylation involved stereoselective enzymatic process.19 Actually, an enzyme, which Tetrahydroisoquinoline derivatives (TIQs) are a group of enantioselectively synthesized (R)-Sal, has been isolated from endogenous neurotoxins, which may be formed from in vivo rat brain.20 condensations of catecholamines with aldehydes.1 Due to the Studies on the stereoselective neurotoxicity of TIQs have similarity in chemical structures, many TIQs have similar promoted the development of analytical methods for the neurochemical properties to 1-[N]-methyl-4-phenyl-1,2,3,6- determination of TIQ enantiomers. Several HPLC procedures21–23 tetrahydropyridine (MPTP), a well-known exogenous and a capillary electrochromatographic (CEC) method24 were causing Parkinsonism in humans, monkeys, and reported for the separation of Sal and N-methyl-Sal enantiomers. various animals.2 MPTP specifically degenerates the Gas chromatography–mass spectrometry (GC-MS) has long nigrostriatal neurons, resulting in dopamine been used for the quantification of catecholic amines.25,26 To deficiency in the striatum, and hence the symptoms of achieve the separation of TIQ enantiomers, Haber et al. developed Parkinsonism. Studies on TIQs neurotoxicity have been a chiral GC-MS method based on a two-step derivatization with extensive, particularly since the discovery of MPTP’s damaging N-methyl-N-trimethylsilyltrifluoracetamide and R-(–)-2- effects on nervous systems.1,3–8 Dopamine-derived TIQs, phenylbutyrylic acid.27 This method was later used by including salsolinol (Sal), have been found to be cytotoxic to rat Musshoff et al. for the determination of dopamine, (R)-/(S)-Sal, PC 12 cells,9 human dopaminergic SH-SY5Y cells10 and and (R)-/(S)-norsalsolinol in human brain tissue samples.28 We melanoma cells.11 There is growing evidence indicating that the reported a chiral GC-MS method employing a β-cyclodextrin in vivo formation of certain TIQs may be related to alcohol (β-CD) coated capillary column for resolving Sal enantiomers.29 addiction.12,13 Although a better resolution for Sal enantiomers was obtained The chirality of TIQs neurotoxicity has been receiving by using the GC-MS method, the water-sensitive derivatization increasing research attention. Many chiral TIQ molecules have procedure involved was not convenient for the assay. Extreme been shown to exhibit enantioselective neurotoxicity.14,15 (S)- care must be taken to obtain reproducible results. Sal was shown to significantly suppress pro-opiomelanocotin The aim of the present work was to develop a chiral capillary (POMC) gene expression, but the (R)-enantiomer did not electrophoresis method to separate catecholamines and their exhibit the same effect.16 After being administered into the derived neurotoxins with high enantiomeric separation striatum, the (R)-enantiomer of N-methyl-Sal induced efficiency. This work also investigated the use of a surfactant to Parkinsonism in rats, while the (S)-enantiomer did not.17 N- improve the recovery rates of these compounds from the Methyl-(R)-Sal was found to be much more potent than its deproteinization procedure widely used for cleaning up optical antipode in inducing apoptosis in dopaminergic biological samples. Finally, the TIQ products formed from in vitro incubation of catecholamines with aldehydes were studied † To whom correspondence should be addressed. using the present method and electrospray ionization mass E-mail: [email protected] spectrometry. 116 ANALYTICAL SCIENCES FEBRUARY 2005, VOL. 21

Experimental

Chemicals β-Cyclodextrin (β-CD), urea, sodium taurocholate (TChl), salsolinol (Sal), N-methyl-(R)-salsolinol (NMSal), dopamine (DA), tetrahydroisoquinoline (TIQ), 1-benzyl-TIQ (BTIQ), epinine, sodium hexanesulfonate (HSA-Na+), Triton X-100, and fetal bovine serum (FBS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Preparation of Sal enantiomers from the racemate was described previously.29 All other reagents used were of analytical grade and Milli-Q water was used throughout. The chemical structures of the important compounds involved in this work are shown in Fig. 1.

Capillary electrophoresis (CE) CE separations were performed on a HP 3D CE instrument (Hewlett-Packard, Palo Alto, CA, USA). Fused-silica capillaries with an effective length of 50 cm (Supelco, Bellefonte, PA, USA) were used. The voltage applied across the capillary was 30 kV. Samples were injected electrokinetically by applying 10 kV for 3 s, unless specified elsewhere. The temperature was maintained at 22˚C. The running buffer consisted of 12 mM β-CD, 1.5 M urea, and 50 mM phosphate buffer (pH 3.0). A diode array detector (DAD) Fig. 1 Chemical structures of certain compounds involved in this was set at 214 nm to monitor the CE eluent. Peak areas were work. used for quantification.

Mass spectrometry (MS) MS measurements were carried out with a Finnigan LCQ were carried out with water replacing the surfactant for a DECA mass spectrometer (Thermo Finnigan, San Jose, CA, comparison. Finally, the solution was diluted with the running USA), which was equipped with an electrospray ionization buffer to an appropriate volume before being injected into CE. (ESI) source operated in the positive-ion mode. Samples were infused into the MS using an attached syringe pump at a flow In vitro incubation studies rate of 5 µl min–1. The operation conditions of the MS detector Epinine was incubated with acetaldehyde under mimic were optimized with a solution of NMSal (500 ng ml–1). The physiological conditions following the procedure described by signal abundance of m/z 194 [M+H]+ was maximized as Kajita et al.30 Briefly, ca. 1 mM epinine, prepared in 50 mM follows: sheath gas flow, 35 arbitrary units; auxiliary gas flow, phosphate buffered saline (PBS) (pH 7.4), was mixed with an 0 arbitrary units; spray voltage, 4 kV; heated capillary excessive quantity of acetaldehyde. The mixture was incubated temperature, 250˚C. A relative collision energy of 30% was at 37˚C for 3 h before being centrifuged at 22000 × g for 10 used for all MS/MS experiments with an isolation width of 3.0 min. The supernatant was collected, and diluted to obtain µm. Other parameters were optimized by the Autotune appropriate concentrations for chiral CE separation and MS program. Nearly identical optimum conditions were found for measurements. Similar incubation tests were also carried out Sal, BTIQ, and NMSal. An Xcalibur software package was using dopamine with acetaldehyde. used to process the data.

Study of the recovery rates of TIQs from deproteinization Results and Discussion DA, TIQ, and (R/S)-Sal were selected as test compounds for these experiments. After 0.3 – 1.0 mg of the test compounds Separation of tetrahydroisoquinoline (TIQ) derivatives were weighed out and mixed with 1.0 ml FBS by vortex, the To the best of our knowledge, there has so far been no report mixture was then left still for at least 1 h at room temperature. on the CE separation of TIQ enantiomers. Therefore, various Procedures using three different precipitants, i.e. methanol CE conditions were studied to achieve optimum enantiomeric

(MeOH), HClO4/MeOH mixture, and trichloroacetic acid separation efficiency. The test compounds included DA, TIQ, (TCA), were tested. When MeOH was used as the precipitant, a BTIQ, Sal, and NMSal. The effects of the capillary inner 200 µl surfactant solution and 400 µl of MeOH were added to diameter were studied using capillaries of 25, 50, and 75 µm i.d. 200 µl of FBS containing DA, TIQ, and Sal. When No significant influence on the separation was observed. A 75

HClO4/MeOH was used, 200 µl of FBS was mixed with 100 µl µm i.d. capillary was selected for an improved reproducibility of a surfactant solution, 200 µl of MeOH and 100 µl of 3 M and a better detection sensitivity, probably due to an increased

HClO4. When TCA was used, 50 µl of surfactant solution and path length. The maximal voltage output available with the 80 µl of a TCA aqueous solution (30%, w/v) were mixed with power supply (30 kV) was used for a shortened running time. 200 µl of FBS. The surfactant solution was either a 3.0 M Various buffer solutions were evaluated for the separation. HSA-Na+ or a 10% (w/v) Triton X-100 aqueous solution. It was Borate, phosphate and acetate buffers were tested with pH added in two different ways: 1) after first being mixed with values ranging from 1.0 to 9.0. It was found that the separation FBS, the precipitant was added; 2) after first being mixed with efficiency was poor at high pH values (i.e. pH > 5.0, borate and the precipitant, the mixture was injected to FBS. Blank tests phosphate buffers). Good separations were achieved in a low ANALYTICAL SCIENCES FEBRUARY 2005, VOL. 21 117

Table 1 Percent recovery of TIQ, dopamine, and (R/S)-Sal from protein precipitation TIQ DA (S)-Sal (R)-Sal Precipitant (No surfactant added) MeOH/HClO4 8.7 ± 2.4 7.0 ± 0.3 6.3 ± 0.6 6.3 ± 1.1 TCA 6.2 ± 1.2 6.6 ± 1.3 5.5 ± 0.0 5.7 ± 0.2 MeOH 11.0 ± 0.8 13.9 ± 1.4 11.3 ± 1.0 11.7 ± 1.0 Precipitant (HAS-Na+ added) a MeOH/HClO4 62.7 ± 3.6 68.7 ± 2.1 56.2 ± 6.4 55.2 ± 0.7 b MeOH/HClO4 16.1 ± 0.5 17.1 ± 3.2 16.3 ± 1.0 14.4v0.2 TCAa 69.0 ± 6.2 51.8 ± 4.4 43.4 ± 5.1 39.7 ± 3.2 TCAb 15.3 ± 2.2 12.4 ± 0.7 11.8 ± 0.1 11.0 ± 0.3 MeOHb 22.2 ± 1.5 22.3 ± 0.9 24.4 ± 2.6 23.5 ± 0.2 Fig. 2 Electropherograms from separation of mixtures of TIQ Precipitant (Triton X-100 added) derivatives. Peak identification: 1, TIQ; 2, dopamine; 3, (S)-Sal; 4, a MeOH/HClO4 61.4 ± 5.3 45.2 ± 2.5 39.8 ± 0.9 37.6 ± 3.3 (R)-Sal; 5, (R)-N-methyl-Sal; 6, (S)-1-benzyl-TIQ; 7, (R)-1-benzyl- b MeOH/HClO4 27.9 ± 3.4 30.7 ± 0.7 19.3 ± 1.7 21.3 ± 4.1 TIQ. “x” indicates an unknown identity. The concentrations of TCAa 58.6 ± 7.7 58.4 ± 6.1 57.3 ± 7.4 52.7 ± 0.5 µ –1 µ –1 µ chemicals were in (a) 50 g ml TIQ, 18 g ml dopamine, 6 g TCAb 33.2 ± 1.3 35.6 ± 4.7 32.4 ± 2.4 32.9 ± 0.6 ml–1 (R)-Sal and (S)-Sal, 10 µg ml–1 (R)-NMSal, and 6 µg ml–1 (R)- and (S)-1-benzyl-TIQ; and in (b) 12.5 µg ml–1 TIQ, 18 µg ml–1 Data shown are means ±S.D. from three determinations. Data with dopamine, (R)-Sal and (S)-Sal, 80 µg ml–1 (R)-NMSal, and 9 µg ml–1 MeOH as a precipitant are not available in all cases because bad (R)- and (S)-1-benzyl-TIQ. The running buffer for (a) contained 12 sedimentation was observed in some cases, and the solution was not mM β-CD, and 50 mM phosphate (pH 3.0); and for (b) was the same run by CE. as (a) except that it also contained 1.5 M urea. a. The detergent was first mixed with FBS, then the precipitant was added. b. The detergent and precipitant were pre-mixed, and were added to FBS solution together. pH range from 1 to 5 using an acetate or a phosphate buffer. A phosphate buffer of pH 3.0 was selected for further experiments. Among various native and modified CDs used as chiral selectors in CE separations, β-CD is the most common, convenient and introduction of a second chiral selector (i.e. sodium economical one. In this work, β-CD alone brought about poor taurocholate, TChl) in addition to β-CD to the running buffer separation results, as shown in Fig. 2(a); none of the further improved the chiral resolution. However, due to the enantiomeric pairs tested could be baseline resolved. The high salt concentration in the running buffer with TChl, the CE addition of urea was found to be very effective for separation in current was almost doubled, which apparently resulted in a poor this case. Urea has been applied to increase the solubility of a baseline. Because the enantiomers were baseline resolved, even CD; it has also been used as an organic modifier,31 and recently without TChl, no further studies were carried out in this for the chiral separation of derivatives.32 As direction, and the running buffer used for further experiments can be seen from Fig. 2(b), TIQ, DA, (R/S)-Sal, (R)-NMSal and did not contain TChl. (R/S)-BTIQ were well separated using a running buffer In this study, the resolution of three pairs of TIQ enantiomers, consisting of 1.5 M urea, 12 mM β-CD, and 50 mM phosphate i.e. (R/S)-BTIQ, (R/S)-Sal, and (R/S)-NMSal were baseline (pH 3.0). No significant change in the separation efficiency was achieved under the CE conditions described above. All (R)- observed when changing the urea concentration within a range enantiomers had a longer migration time than their optical between 0.6 and 3.0 M. The movement of a molecule in CE is antipodes. Elutes can form inclusion complexes with β-CD 31 governed by the electroosmotic (EO) mobility (µeo) and molecules. Theoretical calculations using AM1 semi-empirical 34 electrophoretic mobility (µem), method lead to the conclusion that the (R)-enantiomers of the tested molecules formed more stable inclusion complexes with εζ q β-CD than did their optical antipodes, which indicated that an µ = µeo + µem = —— + ——— (1) η 6πηr (R)-enantiomer would migrate out later. The experimental results were well in accordance with the computation. The where µ represents the mobility of a molecule, ε stands for the calculation results will be published elsewhere. dielectric constant, ζ indicates the zeta potential of the liquid– solid interface and η is the viscosity of the buffer solution; q Recovery studies and r are the net charge and the ionic radius, respectively. For determining the TIQs in biological samples, such as brain Figure 2 shows that in general the migration times of the tissue homogenates, proteins must be removed prior to analysis. analytes increased if urea was added, which could be caused by It was noticed that the recovery rates of TIQ derivatives were an increased viscosity according to Eq. (1). However, in very low after protein precipitation. Pagel et al. reported that comparisons with other chemical species (e.g., TIQ and (R/S)-1- addition of 1-hexanesulfonic acid (HSA) to the sample solution benzyl-TIQ), the migration times for dopamine (peak 2) and before precipitating out proteins could significantly improve the (S)-Sal (peak 3) were almost unchanged, which indicated that recovery rates of some TIQ derivatives.33 In this work, we urea could possibly also work as an organic modifier. It could investigated the influence of two different types of surfactants, have affected other factors in Eq. (1) (such as charge to volume i.e. sodium salt of HSA (HSA-Na+) (negatively charged) and ratios, q/r), and consequently brought about a better resolution. Triton X-100 (non-ionic). The results are listed in Table 1. As It was noticed that the addition of an organic solvent (e.g., can be seen, in the absence of a surfactant the recovery rates methanol, and acetonitrile) into the running buffer diminished were very poor for all of the four test compounds (< 15%). It is the enantiomeric resolution. It is worth noting that the very likely that the analytes were bound to and co-precipitated 118 ANALYTICAL SCIENCES FEBRUARY 2005, VOL. 21

Fig. 3 Electropherogram for recovery study with the addition of HSA-Na+. 1, Dopamine; 2, (S)-Sal; 3, (R)-Sal; “x”, an unknown Fig. 5 Electropherograms obtained from an in vitro identity. See text for details. epinine/acetaldehyde incubation sample (a) and the sample spiked with authentic (R)-NMSal at a concentration of 20 µg ml–1 (b).

In vitro incubation studies To investigate the formation and pathways for TIQ neurotoxins, in vitro incubation tests involving dopamine and epinine (see Fig. 1 for structures) with acetaldehyde were tested. The incubation of epinine with acetaldehyde produced NMSal, a very toxic neurotoxin. Figure 4 shows the results from a MS analysis of an epinine/acetaldehyde incubation sample. As can be seen from Fig. 4(a), the m/z 194 M+1 ion of NMSal (MW = 193) was detected as a predominant ion from the incubation sample. Comparing the MS2 of this ion (Fig. 4(b)) with the MS2 obtained from a solution of authentic (R)- NMSal (Fig. 4(c)) confirmed that the compound was NMSal. Using the present chiral CE separation, the enantiomeric composition of NMSal formed from the incubation was determined. An electropherogram from separating the incubation sample is shown in Fig. 5(a). The separation revealed that NMSal was the predominant product in the sample and equal amounts of (R)-NMSal and (S)-NMSal were formed. To verify the peak identity, the sample was spiked with authentic (R)-NMSal and separated again. The (R)-NMSal peak was confirmed, as can be seen from Fig. 5(b). Similar analyses showed that the incubation of dopamine with acetaldehyde produced equal amounts of (R)-Sal and (S)-Sal, and no NMSal was detected.

Fig. 4 MS analyses of in vitro incubation samples: (a) MS spectrum obtained from an in vitro epinine/acetaldehyde incubation Conclusion sample, (b) MS2 spectrum of m/z 194 ion from (a), and (c) MS2 of m/z 194 ion from an authentic (R)-NMSal sample. This work represents the first chiral capillary electrophoretic (CE) separation of TIQ neurotoxins including Sal, 1-benzyl- TIQ, and N-methyl-Sal. The enantiomeric separation efficiency with serum proteins. If a surfactant was pre-mixed with the was significantly better than those obtained by the chiral GC29 FBS solution, a remarkable improvement in the recovery was or HPLC22 method previously reported. The recovery rates of achieved. Both HAS-Na+ and Triton X-100 showed similar TIQ compounds from deproteinization were investigated. It effects. When a surfactant was pre-mixed with the FBS was found that pre-mixing the sample with a surfactant solution solution, the protein binding sites were largely occupied by (i.e. HAS-Na+ or Triton X-100) significantly improved the surfactant molecules, and the bound analyte molecules were recovery rates. Combined results from MS analysis and chiral released into the aqueous solution. Therefore, analyte/protein CE separation showed that Sal and NMSal were produced from co-precipitation was reduced. It is worth noting that if the in vitro dopamine/acetaldehyde and epinine/acetaldehyde surfactant was pre-mixed with the precipitating reagent instead incubations, respectively. The enantiomeric compositions of of the FBS sample solution, the recovery rates of the analytes Sal and NMSal were also determined to be racemic. Future were not much different from those obtained in the absence of a studies will include the characterization of TIQ neurotoxins surfactant, as shown in Table 1. A typical electropherogram formed from in vitro incubations in the presence of various cell with the addition of HSA-Na+ is shown in Fig. 3. lines and brain tissues. ANALYTICAL SCIENCES FEBRUARY 2005, VOL. 21 119

Melzig, Alcohol, 1995, 12, 447. Acknowledgements 17. M. Naoi, W. Maruyama, Y. Dostert, and Y. Hashizume, J. Neural. Trans. Suppl., 1997, 50, 89. Support of NIH grants (G12RR12459, S06GM08047 and 18. W. Maruyama, A. A. Boulton, B. A. Davis, P. Dostert, and NS44177) is gratefully acknowledged. M. Naoi, J. Neural Transm., 2001, 108, 11. 19. Th. Muller, S. S. Baum, D. Haussermann, D. Woitalla, H. Rommelspacher, H. Przuntek, and W. Kuhn, J. Neural References Transm., 1998, 105, 239. 20. M. Naoi, W. Maruyama, P. Dostert, K. Kohda, and T. 1. T. Nagatsu, Neurosci. Res., 1997, 29, 99. Kaiya, Neurosci. Lett., 1996, 103, 183. 2. R. E. Heikkila, A. Hess, and R. C. Duvoisin, Science, 1984, 21. S. S. Baum and H. Rommelspacher, J. Chromatogr. B, 224, 1451. 1994, 660, 235. 3. J. Vetulani, L. Antkiewicz-Michaluk, I. Nalepa, and M. 22. Y. Deng, W. Maruyama, M. Kawai, P. Dostert, H. Sansone, Neurotox. Res., 2003, 5, 147. Yamamura, T. Takahashi, and M. Naoi, J. Chromatogr. B, 4. L. Antkiewicz-Michaluk, I. Romanska, I. Papla, J. 1997, 689, 313. Michaluk, M. Bakalarz, J. Vetulani, A. Krygowska-Wajs, 23. K. McMurtrey, C. Strawbridge, and J. G. McCoy, and A. Szczudlik, Neuroscience, 2000, 96, 59. Enantiomers, 2000, 5, 377. 5. A. Thumen, A. Behnecke, F. Qadri, E. Bauml, A. Thumen, 24. Y. Deng, J. Zhang, T. Tsuda, P. H. Yu, A. A. Boulton, and C. A. Behnecke, F. Qadri, E. Bauml, and A. Moser, R. M. Cassidy, Anal. Chem., 1998, 70, 4586. Neuroreport, 2002, 13, 25. 25. T. Niwa, W. Maruyama, D. Nakahara, N. Takeda, H. 6. A. Storch, S. Ott, Y. I. Hwang, R. Ortmann, A. Hein, S. Yoshimizu, A. Tatematsu, A. Takahashi, P. Dostert, M. Frenzel, K. Matsubara, S. Ohta, H. U. Wolf, and J. Naoi, and T. Nagatsu, J. Chromatogr., 1992, 578, 109. Schwarz, Biochem. Pharmacol., 2002, 63, 909. 26. H. Haber, H. Haber, and M. F. Melzig, Anal. Biochem., 7. S. Shavali and M. Ebadi, Neurotoxicology, 2003, 24, 417. 1995, 224, 256. 8. Y. Kotake, S. Ohta, I. Kanazawa, and M. Sakurai, 27. H. Haber, P. Henklein, M. Georgi, and M. F. Melzig, J. Neuroscience, 2003, 117, 63. Chromatogr. B, 1995, 672, 179. 9. Y. J. Surh, Y. J. Jung, J. H. Jang, J. S. Lee, and H. R. Yoon, 28. F. Musshoff, P. Schmidt, R. Dettmeyer, F. Priemer, K. J. Toxicol. Environ. Health A, 2002, 65, 473. Jachau, and B. Madea, Forensic Sci. Int., 2000, 113, 359. 10. S. Shavali, J. Ren, and M. Ebadi, Neurosci. Lett., 2003, 29. Y.-M. Liu, P. Gordon, S. Green, and J. Sweedler, Anal. 340, 79. Chim. Acta, 2000, 420, 81. 11. F. De Marco, M. Perluigi, M. L. Marcante, R. Coccia, C. 30. M. Kajita, T. Niwa, W. Maruyama, D. Nakahara, N. Foppoli, C. Blarzino, and M. A. Rosei, Biochem. Takeda, H. Yoshizumi, A. Tatematsu, K. Watanabe, M. Pharmacol., 2002, 64, 1503. Naoi, and T. Nagatus, J. Chromatogr. B, 1994, 654, 263. 12. R. D. Myers, Alcohol Clin. Exp. Res., 1996, 20, 498. 31. S. Ahuja, “Chiral Separations by Chromatography”, 2000, 13. M. Jamal, K. Ameno, T. Kubota, S. Ameno, X. Zhang, M. Oxford University Press, New York. Kumihashi, and I. Ijiri, Alcohol Alcohol., 2003, 38, 197. 32. Y. S. Huang, J. T. Liu, L.-C. Lin, and C.-H. Lin, 14. K. Abe, K. Taguchi, T. Wasai, J. Ren, I. Utsunomiya, T. Electrophoresis, 2003, 24, 1097. Shinohara, T. Miyatake, and T. Sano, Brain Res. Bull., 33. P. Pagel, R. Schubert, and H. U. Wolf, J. Chromatogr. B, 2001, 56, 55. 2000, 746, 283. 15. M. Naoi, W. Maruyama, Y. Akao, and H. Yi, Neurotoxicol. 34. R. Weinberger, “Practical Capillary Electrophoresis”, 2nd Teratol., 2002, 24, 579. ed., 2000, Academic Press, San Diego. 16. I. Putscher, H. Haber, A. Winkler, J. Fickel, and M. F.