Separation and Purification of Methadone Enantiomers by Continuous

using a Model IPS-16 peristaltic pump (Ismatech, München, FRG;. 3 in Figure 2), and buffer flow ... into the chamber (five inlet ports at bottom of F...
0 downloads 0 Views 145KB Size
Download PDF
Anal. Chem. 1999, 71, 1840-1850

Separation and Purification of Methadone Enantiomers by Continuous- and Interval-Flow Electrophoresis Peter Hoffmann,† Horst Wagner,† Gerhard Weber,‡ Matthias Lanz,§ Jitka Caslavska,§ and Wolfgang Thormann*,§

Institut fu¨ r Anorganische und Analytische Chemie und Radiochemie, Fachrichtung 11.4 der Universita¨ t des Saarlandes, Postfach 151150, D-66041 Saarbru¨ cken, Germany, Dr. Weber GmbH, Klausnerring 17, D-85551 Kirchheim, Germany, and Department of Clinical Pharmacology, University of Bern, Murtenstrasse 35, CH-3010 Bern, Switzerland

About one-fourth of all therapeutic agents are administered to man as mixtures of isomeric substances whose biological activity may well reside predominantly in one form. The use of racemic mixtures typically results in stereoselective drug metabolism and may also contribute to the toxicity or adverse effects encountered with drugs. Not surprisingly, administration of pure drug enantiomers is often advantageous for pharmacotherapy.1 Furthermore, systematic investigation of the biological activity (including pharmacology and toxicology) of individual enantiomers became

the rule not only for all new racemic drugs but also for new racemic agrochemicals. Thus, separation of enantiomers is an important topic in the pharmaceutical and agrochemical industries There are two approaches for obtaining enantiomerically pure substances. These are (i) asymmetric synthesis of the desired isomer and (ii) synthesis of both compounds and resolution of the racemic mixture into individual isomers. Asymmetric synthesis is time-consuming and expensive and is thus justified only when a large quantity of an enantiomer is required. Synthesis of a racemic mixture is simpler, and after separation, both enantiomers are available for testing purposes. The production of mass quantities of racemates of commercial importance has created a need for large-scale purification techniques, and if a particular product is to be used clinically, a high degree of purity is essential. Commonly used purification methods for enantiomers include recrystallization and chromatography, the latter approach being more widely applied, namely via formation of diastereomers or via direct separation on chiral stationary phases.2,3 Like all methods, the chromatographic approaches suffer from certain drawbacks, including high costs of stationary phases, necessary dilution of the sample, consumption of large amounts of mobile phase, and difficulties in reusing the mobile phase. In addition, most chromatographic procedures are necessarily batch, as opposed to continuous, procedures, and it is difficult to monitor and control a separation while it is in progress. Various approaches have been employed for improving chromatographic throughput and resolution, such as close injections, peak shaving, peak recycling,4 and application of a process termed simulated-movingbed chromatography.5,6 The latter approach is reported to provide cost-effective large-scale separations with high preservation of the mobile phase. Its instrumental complexity, however, has thus far prevented a widespread adoption of this technology for separation and purification of enantiomers. In the past two decades, other separation techniques have come under scrutiny for application to large-scale fractionations, including continuous- and recycling-flow electrophoresis (two free-

* Corresponding author: (phone) 41 31 632 3288, (fax) 0 41 31 632 4997, (email) [email protected]. † Universita ¨t des Saarlandes. ‡ Dr. Weber GmbH. § University of Bern. (1) Caldwell, J. J. Chromatogr., A 1995, 694, 39-48.

(2) Francotte, E. J. Chromatogr., A 1994, 666, 565-601. (3) Francotte, E. R. Chimia 1997, 51, 717-725. (4) Dingenen, J.; Kinkel, J. N. J. Chromatogr., A 1994, 666, 627-650. (5) Cavoy, E.; Deltent, M.-F.; Lehoucq, S.; Miggiano, D. J. Chromatogr., A 1997, 769, 49-57. (6) Francotte, E. R.; Richert, P. J. Chromatogr., A 1997, 769, 49-57.

Continuous- or free-flow electrophoresis is based upon a thin film of fluid flowing between two parallel plates. The electrolytes and the sample are continuously admitted at one end of the electrophoresis chamber and are fractionated by an array of outlet tubes at the other. Using the Octopus apparatus in a horizontal position, continuous preparative separation of methadone enantiomers in the presence of (2-hydroxypropyl)-β-cyclodextrin as a chiral selector was investigated under conditions of continuousflow zone electrophoresis and continuous-flow isotachophoresis. The enantiomeric composition of methadone in the collected fractions was assessed by chiral capillary electrophoresis and circular-dichroism spectroscopy. In both electrophoretic modes, partial separation of the two enantiomers with an enrichment of about 80% and a throughput of 10-20 mg of racemic methadone per hour was obtained. Operating the Octopus apparatus with interrupted buffer flow during electrophoresis, a process termed interval-flow electrophoresis, resulted in complete separation of milligram quantities of the two methadone enantiomers. Furthermore, commencing with racemic methadone, continuous multistage isotachophoretic processing is shown to be suitable to purify (R)-(-)-methadone, the enantiomer with higher pharmacological activity, on a mg/h scale and at a mM concentration in the collected product stream.

1840 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

10.1021/ac981178v CCC: $18.00

© 1999 American Chemical Society Published on Web 04/02/1999

fluid approaches) as well as electrophoresis in solid-support media.7-9 Although these approaches have primarily been applied to the purification and isolation of proteins and cells or cell organelles, two recent papers discuss the use of preparative electrophoresis for the chiral separation of milligram quantities of drug enantiomers. Stalcup et al. reported gel electrophoretic approaches for separation of terbutaline enantiomers,10 and Lanz et al. discussed the use of recycling free-fluid isotachophoresis (ITP) for the separation of methadone enantiomers.11 In analogy to analytical electromigration methods, which have been studied extensively and shown to provide high resolution at low cost,12-16 enantiomeric separation in both reported approaches is based upon the presence of a chiral selector that has been added to the running buffer (sulfated cyclodextrin) and leading electrolyte ((2hydroxypropyl)-β-cyclodextrin (OHP-β-CD)), respectively. The enantiomeric separation data reported indicate that gel electrophoresis provides higher resolution as compared with that obtained through processing in free solution. This does not come as a surprise, as the gel matrix is known to stabilize the fluid against hydrodynamic (Poiseuille flow), electroosmotic, and electrohydrodynamic (convection originating at conductivity and dielectric gradients) disturbances, phenomena that are known to limit resolution in free-fluid electrophoresis.9,17,18 On the other hand, the throughput in the gel format is low.9,10 Furthermore, ITP as opposed to zone electrophoresis (ZE) has the advantage of regulating the concentration of the processing solution.11 Alternatively, enantiomer resolution in immobilized pH gradient gels via inclusion of a chiral selector has been demonstrated and proposed for exploitation of enantiomeric separations in multicompartment electrolyzers with isoelectric membranes.19 However, no paper was found in the Literature describing the purification of an enantiomer from a racemate using this technology. Our laboratories have many years of experience in preparative free-fluid continuous-flow20-22 and recycling flow9,11,23,24 electrophoretic technology. These processes are based upon a thin film of fluid flowing between two parallel plates. In continuous-flow (7) Bier, M. Electrophoresis 1998, 19, 1057-1063. (8) Krˇiva´nkova´, L.; Bocˇek, P. Electrophoresis 1998, 19, 1064-1074. (9) Thormann, W. In Protein Purification, Principles, High-Resolution Methods, and Applications; Janson, J.-C., Ride´n, L., Eds; Wiley-VCH: New York, 1998; pp 651-678. (10) Stalcup, A. M.; Gahm, K. H.; Gratz, S. R.; Sutton, R. M. C. Anal. Chem. 1998, 70, 144-148. (11) Lanz, M.; Caslavska, J.; Thormann, W. Electrophoresis 1998, 19, 10811090. (12) Snopek, J.; Jelı´nek, I.; Smolkova´-Keulemansova´, E. J. Chromatogr. 1992, 609, 1-17. (13) Novotny, M.; Soini, H.; Stefansson, M. Anal. Chem. 1994, 66, 646A-655A. (14) Ward, T. J. Anal. Chem. 1994, 66, 633A-640A. (15) Fanali, S. J. Chromatogr., A 1996, 735, 77-121. (16) Nishi, H.; Terabe, S. J. Chromatogr., A 1995, 694, 245-276. (17) Rhodes, P. H.; Snyder, R. S.; Roberts, G. O. J. Colloid Interface Sci. 1989, 129, 78-90. (18) Roberts, G. O.; Rhodes, P. H.; Snyder, R. S. J. Chromatogr. 1989, 480, 3567. (19) Righetti, P. G.; Ettori, C.; Chafey, P.; Wahrmann, J. P. Electrophoresis 1990, 11, 1-4. (20) Wagner, H.; Kessler, R. GIT Labor. Med. 1984, 7, 30-35. (21) Wagner, H. Nature (London) 1989, 341, 669-670. (22) Keuth, U.; Leinenbach, A.; Beck, H. P.; Wagner, H. Electrophoresis 1998, 19, 1091-1096. (23) Caslavska, J.; Gebauer, P.; Thormann, W. Electrophoresis 1994, 15, 11671175. (24) Caslavska, J.; Thormann, W. Electrophoresis 1994, 15, 1176-1185.

(or free-flow) electrophoresis, the electrolytes and the sample are continuously admitted at one end of the electrophoresis chamber and are fractionated by an array of outlet tubes at the other end. In recycling electrophoresis, the effluent from each channel is continuously reinjected into the cell through the corresponding influent port, and fluid recycling is maintained until final separation is attained. This latter approach has been recently applied to the separation of methadone enantiomers.11 Using ITP with a counterflow of leading electrolyte and OHP-β-CD as chiral selector, (R)-(-)-methadone and (S)-(+)-methadone were found to become enriched significantly (up to about 80%) at the front and backside, respectively, of the isotachophoretic sample train. To our knowledge, continuous-flow electrophoresis techniques have not yet been applied to the separation and purification of enantiomers. The main advantage of this processing format is the continuous character as opposed to the batch approaches discussed thus far.10,11 This paper reports the first investigation focusing on enantiomeric separation and purification by continuous-flow electrophoresis. Methadone (for structure see Figure 1A), whose enantiomers can be easily separated by chiral capillary zone electrophoresis (CZE) and chiral capillary ITP (Figure 111), was selected as the model compound. It represents an analgesic drug with pharmacological properties similar to those of morphine. It is clinically used for the treatment of severe pain and the chronic maintenance treatment of heroin addiction. It is well-known that the two methadone enantiomers show different pharmacodynamic and pharmacokinetic parameters, including protein binding.25,26 The R form is the enantiomer with higher pharmacological activity and is thus prescribed as a single-isomer preparation in Germany. In many countries, including Switzerland, however, the racemate is registered and administered. Not surprisingly, methadone is characterized by a stereoselective metabolism.27,28 Using the Octopus continuous-flow apparatus, continuous-flow preparative separation of methadone enantiomers in the presence of OHP-βCD was investigated under conditions of ZE and ITP, and the isolation of the pharmacologically more potent (R)-(-)-methadone enantiomer by isotachophoretic multistage processing was studied. Furthermore, separation of enantiomers was also characterized using interval-flow ZE and ITP, processes in which the imposed flow was interrupted during electrokinetic separation.29 The enantiomeric composition of methadone in the collected fractions was assessed by chiral CZE and circular-dichroism spectroscopy. EXPERIMENTAL SECTION Chemicals and Standard Solutions. All chemicals used were of analytical or research grade. Racemic methadone-HCl of European Pharmacopoeia quality was from the hospital pharmacy (Inselspital, Bern, Switzerland). Pure standards of (S)-methadone and (R)-methadone were kindly received from Dr. P. Baumann (25) Moffat, A. C., Ed. Clarke’s Isolation and Identification of Drugs in pharmaceuticals, body fluids and post-mortem material, 2nd ed.; The Pharmaceutical Press: London, 1986; pp 742-743. (26) Eap, C. B.; Cuendet, C.; Baumann, P. Clin. Pharmacol. Ther. 1990, 47, 338-346. (27) Nakamura, K.; Hachey, D. L.; Kreek, M. J.; Irving, C. S.; Klein, P. D. J. Pharm. Sci. 1982, 71, 40-43. (28) Lanz, M.; Thormann, W. Electrophoresis 1996, 17, 1945-1949. (29) Bauer, J.; Weber, G. J. Dispersion Sci. Technol. 1998, 19, 937-950.

Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1841

Figure 2. Schematic representation of the continuous-flow electrophoresis apparatus used for continuous-flow zone electrophoresis. Key: 1, separation chamber with a chamber width and electrode length denoted by a and b, respectively; 2, buffer reservoir; 3, buffer pump; 4, electrode buffer circulation; 5, fraction collector; 6, anodic electrode compartment; 6′, cathodic electrode compartment; 7, spacer between front and back plates (gray); 8, sample pump; 9, counter flow for fractionation.

Figure 1. Enantiomeric separation of racemic methadone obtained in CZE at pH 2.5 (BioFocus 3000 with 195-nm detection and 20 mg/ mL sample, see Experimental Section) and cationic capillary ITP with a leader pH of 4.3 (Tachophor with conductivity detector and 10 mM sample, for details refer to ref 11). The asterisk marks the chiral center of methadone. The y-scale response in panel B expresses the increase in resistance within the isotachophoretic zones. L and T refer to leader and terminator, respectively.

of the Department of Adult Psychiatry, University of Lausanne (Hoˆpital de Ce´ry, Prilly-Lausanne, Switzerland). OHP-β-CD (degree of substitution ∼0.6) was from Fluka (Buchs, Switzerland). Continuous-Flow Instrumentation and Operation Conditions. An overall schematic representation of the instrumental setup used for continuous-flow ZE is given in Figure 2. In that operational mode, electrolytes and the sample are continuously admitted at one end of the electrophoresis chamber and are fractionated by an array of outlet tubes at the other end. Continuous-flow experiments were performed on an Octopus apparatus (Dr. Weber GmbH, Kirchheim, FRG) having chamber dimensions for length, width, and fluid-layer thickness of 500, 100, and 0.4 mm, respectively, an electrode length of 450 mm (distance b in Figure 2), and 96 equidistant outlet ports. The back plate comprises a thermostated aluminum slab that is covered by a glass plate with a thin silver layer (mirror) on the backside, whereas the front plate is made from plexiglass. The latter part comprises all inlet and outlet ports. A 0.4-mm PVC spacer was used to define 1842 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

the gap between the two plates. The instrument was operated in the horizontal position, and the electrode chambers were separated from the separation chamber by PP60 membranes (Dr. Weber GmbH). Buffer and counterflow were continuously infused using a Model IPS-16 peristaltic pump (Ismatech, Mu¨nchen, FRG; 3 in Figure 2), and buffer flow through the electrode chambers (6,6′ in Figure 2) was maintained by a membrane pump model MD-100 (KNF-Neuberger, Freiburg, FRG; 4 in Figure 2). Sample was continuously infused through port 8 using a peristaltic pump (Dr. Weber GmbH). Fractions (about 1 mL) were collected into 96 plastic test tubes of 10-mL volume each. The power supply was a Model 2197 (LKB, Bromma, Sweden). The cooling temperature of the recirculation fluid was maintained at 5 °C. The Octopus was operated in two electrophoretic modes. For ZE, the separation buffer containing the chiral selector was infused into the chamber (five inlet ports at bottom of Figure 2) and the fraction collector (line 9 of Figure 2) from reservoir 2 by pump 3, and both electrode chambers were continuously flushed with 0.1 M KH2PO4. For ITP, leader containing the chiral selector was infused through ports 2-5 (numbering of inlet ports from anode to cathode) and into the fraction collector (line 9 of Figure 2), whereas terminator (without chiral selector) was introduced through port 1 (left-hand side inlet on bottom of Figure 2). The anolyte was composed of 1 M acetic acid and the catholyte was 100 mM acetic acid/sodium acetate (pH 4.34). Instrumentation and Operation Conditions Used for Interval-Flow Electrophoresis. The setup used for interval-flow electrophoresis was similar to that employed for continuous-flow experiments (Figure 2, cf. above). The chamber dimensions were 500, 140, and 0.4 mm; there were 9 inlet ports for buffer (bottom); the sample inlet was located between ports 2 and 3; the fractions (200-250 µL each) were collected into 134 wells of two microtiter plates; and the cell was operated horizontally. Interval-flow

electrophoresis was operated with buffers and sample containing 0.2% (w/v) hydroxypropylmethylcellulose (HPMC). The recirculating cooling fluid was maintained at 8 °C. In interval-flow electrophoresis, electrolytes and the sample were introduced without application of electric power. The sample contained an anionic dye which permitted a visual alignment of the sample band with length b (Figure 2) in the part of the cell that housed the driving electrodes. Before and after sample was introduced, buffer or a solution of equal conductivity (10 mM HCl/glutamine) was introduced through the sample-inlet port. This arrangement prevented the sample at the two edges (absence of electrodes) from being distorted upon current application. After full alignment of sample and buffers along the entire cell, pump 3 was inactivated and electric power applied. For fractionation, the power supply was disconnected and pump 3 was again activated; the first few drops were discarded prior to collection of the content of the separation chamber (cell volume along distance b of Figure 2). For interval-flow ZE, the separation buffer containing the chiral selector was infused into the chamber through ports 2-8 and into the fraction collector from reservoir 2 by pump 3 (Figure 2). A solution composed of 70 mM phosphoric acid and 100 mM glutamine was infused through port 1 (anode), and one of 100 mM acetic acid/100 mM sodium acetate was introduced through port 9 (cathode). Electrode chambers were continuously flushed with 0.1 M H3PO4 (anolyte) and 100 mM acetic acid/100 mM sodium acetate (pH 4.5, catholyte). For interval-flow ITP, leader containing the chiral selector was infused through ports 3-8, 10-fold more concentrated leader through port 9, and 0.2 M acetic acid through ports 1 and 2 (Figure 2). The solution infused into port 2 also contained some low molecular mass spacers (GABA, piperidine carbonic acid, β-alanine, EACA, creatinine, melanine (8 mM each)). The counter flow introduced into the fractionator comprised leader without chiral selector. The anolyte was composed of 1 M acetic acid and the catholyte was 100 mM acetic acid/100 mM sodium acetate (pH 4.5). Analysis of Collected Fractions. For pH measurement, a pH meter, model 720, and a ROSS pH electrode, model 8103 (both from Orion Research, Cambridge, MA) were used. The conductivity was measured with a conductivity meter model 101 (Orion Research, Cambridge, MA) equipped with a model PW 9510/65 cell (Philips, Eindhoven, The Netherlands). The absorbance (determined in 100-fold diluted fractions) was measured at 206 nm using a UV-VIS spectrophotometer Lambda 15 (Perkin-Elmer, Ueberlingen, Germany). Fractions were also analyzed by circulardichroism spectroscopy. Circular-dichroism spectra (ellipticity in millidegree (mdeg) vs wavelength) were recorded on a model J-710 spectropolarimeter (Jasco, Tokyo, Japan) and at room temperature using a cell of 1-mm optical path length. The wavelength was scanned at 20 nm/min between 330 and 250 nm, the bandwidth was 1 nm and the resolution 0.5 nm. The method used for the CZE analysis of the enantiomeric composition of the fractions was the same as that employed previously using a pH 2.5 buffer composed of 50 mM KH2PO4 containing 5 mM OHP-β-CD.11,28 Experiments were performed on a BioFocus 3000 capillary electrophoresis system (Bio-Rad Laboratories, Hercules, CA) that was equipped with an untreated 50 µm i.d. fused-silica capillary (Polymicro Technologies, Phoenix,

Figure 3. Chiral continuous-flow ITP data obtained with a leader composed of 10 mM sodium acetate/acetic acid (pH 4.3) and 5 mM OHP-β-CD. UV absorbance (206 nm), pH, and conductivity data of the collected fractions are shown in panel A. Conductivity data are in mS/cm and are altered by 2x + 2.2 (x ) measured conductivity) for presentation purposes. Panel B shows the enantiomeric composition of the fractions as assessed by chiral CZE. Chiral CZE data of fractions 68, 72, and 77 (denoted as F68, F72, and F77, respectively) are presented in panel C. For CZE analysis, the BioFocus 3000 was employed, and fractions were diluted 10-fold with water. Experimental conditions were as stated in the Experimental Section.

AZ) of 24 cm in total length (19.4 cm to the detector) that was mounted in a user-assembled cartridge (Bio-Rad). Injection of sample was effected by applying a positive pressure (2 psi ‚ s). A constant voltage of 12 kV (current about 54 µA) was applied, the temperatures of cartridge and carousels were maintained at 20 °C, and detection was effected at 195 nm. Before each experiment the capillary was rinsed with 0.1 M NaOH for 1 min, water for 0.5 min, and running buffer for 1.5 min. Most of the fractions were diluted (up to 100 times) prior to analysis. With that assay, run times of about 5 min were obtained, and the peak resolution Rs was better than 4 (Figure 1). BioFocus integration software (version 5.0, Bio-Rad) was employed for data conversion and evaluation. Data were expressed in relative peak areas, i.e., the peak area divided by its detection time. Calibration graphs constructed from four concentration levels (between 2.53 and 10.11 µg/mL of each enantiomer) were found to be linear (y ) 1.700 × 10-4x - 0.023 and y ) 1.696 × 10-4x - 0.016 for (R)-(-)methadone and (S)-(+)-methadone, respectively) with F values > 1000 (P < 0.001; r > 0.9996). Alternatively, fractions were also analyzed by the same chiral assay using the HP 3D capillary electrophoresis system (Hewlett-Packard, Waldbronn, Germany) that was equipped with a bare fused-silica capillary of 50 µm i.d. Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1843

Table 1. Selected Chiral ITP Separation Data for Methadonea continuous-flow ITP fraction no.

R-METb (rel area %)

64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80

13.3 11.4 9.10 10.1 17.2 18.3 39.0 44.0 49.2 55.2 60.8 66.9 74.6 80.0 79.4 73.6 75.8

ee R-MET (%)

interval-flow ITP ee S-MET (%)

fraction no.

R-METb,c (rel area %)

73.4 77.3 81.8 79.9 65.6 63.5 22.0 12.0 1.70

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

0.00 0.00 1.23 1.89 5.04 7.19 19.3 33.6 55.6 86.1 100 100 100 100 100

10.4 21.5 33.9 49.2 60.0 58.9 47.3 51.6

ee R-METc (%)

ee S-METc (%) 100 100 97.5 96.2 89.9 85.6 61.3 32.7

11.2 72.3 100 100 100 100 100

a Data presented correspond to those of Figures 3 (continuous-flow) and 5 (interval-flow). b Corresponding S-MET data could be calculated by 100% - (R-MET). c 100% refers to the case in which the second enantiomer could not be detected (>99.7% purity for 1 mM concentration).

Figure 4. Circular-dichroism data obtained with (A) standard solutions and (B) fractions 68-75 (denoted as F68-F75, respectively) of the run of Figure 3. Circular-dichroism measurements were made as described in the Experimental Section.

capillary of 30 cm (38.5 cm) effective (total) length. Nineteen kilovolts (ca. 60 µA) was applied, chamber air was thermostated at 25 °C, pressure injection of sample was effected at 5 mbar for 3 s, and detection was performed at 195 nm. Quantitation was executed via seven-level calibration in the range between 0.0125 and 1.0 mg/mL of racemic methadone. Calibration graphs were found to be linear (y ) 544.6 x - 0.231 and y ) 418.3 x - 14.82 for ZE and ITP, respectively) with F values > 250 (P < 0.001; r ) 0.9996 (M in ZE buffer); r ) 0.996 (M in ITP leader)). The detection limit (S/N ) 3) was determined to be 1 µg/mL (3.24 µM). Thus, at a 1 mM drug level, enantiomeric purity > 99.7% could not be elucidated. 1844 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

RESULTS AND DISCUSSION Separation of Methadone Enantiomers by Chiral Capillary Zone Electrophoresis and Isotachophoresis. Methadone has a pKa value of about 8.3 and is thus positively charged up to about pH 10.3. Enantiomeric separation was reported to require a buffer of low pH.28 Using a 50 mM KH2PO4 buffer at pH 2.5 containing 5 mM OHP-β-CD, the enantiomers were found to separate well (Figure 1A). Enantiomeric resolution (Rs) can be characterized by Rs ) 1.18(t2 - t1)/((Wh/2)1 + (Wh/2)2) where ti and (Wh/2)i represent the detection time and peak width at half-height of enantiomer i, respectively. First- and second-detected enantiomers

Figure 5. Chiral interval-flow ITP data obtained with a pH 4.5 leader containing 5 mM OHP-β-CD as chiral selector. Panel A shows the enantiomeric composition of the fractions as assessed by chiral CZE and panel B depicts electropherograms of selected undiluted fractions that were monitored with the HP instrument. Other conditions were as in the text.

are denoted with subscripts 1 and 2, respectively. An Rs value of g1.4 represents baseline resolution. In the presented example, the peak resolution Rs was determined to be better than 4. Furthermore, enantiomeric separation depends on the difference of the electrophoretic mobilities of the two enantiomers, a quantity that can be calculated according to ∆m ) Leff/E((t2 - t1)/(t1t2)) where Leff and E represent the effective column length and the applied electric field, respectively. For the data presented in Figure 1A, the mobility difference was estimated to be 1.08 × 10-5 cm2/ (V s). Similarly, employing the ZE buffer composed of 63 mM acetic acid and 7.0 mM L-glutamine (pH 3, κ ) 0.35 mS/cm) containing 5 mM OHP-β-CD as chiral selector (see below) provided data with ∆m ) 1.3 × 10-5 cm2/(V s) (data not shown). As discussed previously, methadone behaves isotachophoretically having sodium or potassium as leading compound, acetate as counter constituent, and H3O+ (acetic acid) as terminator. Good separations of enantiomers were noted for leader pH values between 4.0 and 4.8 and having 5-20 mM of OHP-β-CD in the leader.11 An isotachopherogram illustrating the separation of the two enantiomers is shown in panel B of Figure 1. For that separation, the mobility difference between the two enantiomers at steady state was estimated to be about 5 × 10-5 cm2/(V s). Continuous- and Interval-Flow Isotachophoresis of Methadone Enantiomers. The data presented in Figure 3 were

obtained by chiral continuous-flow ITP having 10 mM sodium acetate/acetic acid (pH 4.3) with 5 mM of the chiral selector as leader. The terminator was 10 mM acetic acid. The UV absorbance, pH, and conductivity distributions obtained are shown in panel A, whereas panel B depicts the enantiomeric composition of the fractions as assessed by chiral CZE (expressed in relativepeak-area values (see below); for examples of electropherograms see Figure 3C). In that run, the residence time was 5 min, the voltage was kept constant at 600 V, and equilibrium was attained after 25 min of power application. From that time point on, the current was 32 mA (applied electric power, 19.2 W), and fraction collection was executed during about 20 min. The rate of counterflow of leader (line 9 of Figure 2) used was about 1.2-fold higher as compared with the total flow of 4 mL/min through the cell. Thus, at the collection outline ports, the chamber fluid was continuously mixed with leader (about 2.2-fold dilution with leader), and the measured properties of the collected fractions outside the isotachophoretic leader region (fraction numbers < 79) were correctly found to be different than those computer predicted and measured in recycling ITP.11 Nevertheless, a plateau-shaped region (fractions 69-76) containing methadone was obtained (Figure 3A), the methadone concentration being about 2.3 mM, a value that is about 2.2-fold smaller compared with that obtained without dilution during sample collection.11 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1845

Figure 6. Chiral continuous-flow ZE data at pH 3 with 5 mM OHP-β-CD as the chiral selector. Panel A shows the enantiomeric composition of the fractions as assessed by chiral CZE, and panel B depicts electropherograms of selected fractions that were diluted 5-fold with water and monitored with the BioFocus 3000.

Thus, methadone which was infused at a concentration of 10 mg/ mL (32.3 mM) becomes diluted about 14-fold (6.46-fold dilution comes from the ITP adjustment). Fractions 1-60 and 80-96 did not contain any measurable methadone, whereas small amounts of methadone were monitored in fractions 61-68 and 77-79. Chiral separation could not be recognized via analysis of the fractions by UV absorption, conductivity, or pH (Figure 3A). However, chiral CZE analysis of the fractions revealed that there is an enrichment of (R)-(-)-methadone at the ITP front, an almost racemic mixture of the two enantiomers in the center of the methadone zone, and a clear enrichment of (S)-(+)-methadone at the rear end. This is nicely seen with the relative-peak-area data presented in Figure 3B and the R-MET data presented in Table 1. The R-MET data (expressed as relative area %) were obtained according to (R)-MET ) (relative peak area of (R)(-)-methadone/(relative peak area of (R)-(-)-methadone + relative peak area of (S)-(+)-methadone)) ‚ 100%. Furthermore, in analogy to the enantiomeric excess that is typically defined as (difference of enantiomer concentrations/sum of enantiomer concentrations)‚100%,30 an enantiomeric excess, ee, was calculated (30) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds; John Wiley & Sons: New York, 1994.

1846 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

according to

ee ) (difference in relative peak areas/ sum of relative peak areas) × 100%

and obtained values are listed in Table 1. The maximum ee values for (S)-(+)-methadone and (R)-(-)-methadone were determined to be 81.8% (fraction 66) and 60.0% (fraction 77), respectively. ITP zone formation and tailing are in agreement with the observations made in analytical and recycling ITP.11 As was the case with recycling ITP, circular-dichroism spectroscopy was used for characterization of the fractions (Figure 4). With that method, the difference in absorbance of left and right circularly polarized light of the two enantiomers as a function of the wavelength is monitored. Although the chiral selector also exhibits optical activity, it did not affect the circular-dichroism measurement because it does not show absorbance in the wavelength range used. The spectrum obtained with an aqueous solution of racemic methadone did not reveal any signal in the scanned wavelength range between 330 and 250 nm.11 In the presence of the chiral selector, however, a small positive response was observed (Figure 4A). This solute-induced circular dichroism

Table 2. Selected Chiral ZE Separation Data for Methadonea continuous-flow ZE fraction no.

R-METb (rel area %)

53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93

16.7 16.4 26.4 32.4 37.4 40.0 42.5 44.0 45.6 47.0 49.0 51.1 53.4 56.5 59.9 64.1 69.0 75.9 81.4 86.3 90.1

ee R-MET (%)

interval-flow ZE ee S-MET (%)

fraction no.

R-METb,c (rel area %)

66.7 67.1 47.3 35.2 25.2 19.9 15.0 12.0 8.80 6.00 2.00

70 72 74 76 78 80 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

0 0 0 0 0 0 0.91 2.03 7.33 20.2 34.6 45.0 52.8 58.2 78.6 96.7 100 100 100 100 100

2.10 6.90 12.9 19.7 28.2 38.0 51.8 62.8 72.7 81.4

ee R-METc (%)

ee S-METc (%) 100 100 100 100 100 100 98.2 95.9 85.4 59.6 30.8 10.1

5.66 16.5 57.3 93.3 100 100 100 100 100

a Data presented correspond to those of Figures 6 (continuous-flow) and 7 (interval-flow). b Corresponding S-MET data could be calculated by 100% - (R-MET). c 100% refers to the case in which the second enantiomer could not be detected (>99.7% purity for 1 mM concentration).

is in agreement with data published by Han et al.31 Furthermore, measuring solutions containing (R)-(-)-methadone and (S)-(+)methadone revealed negative and positive peaks, respectively, with minimum/maximum responses at 295 nm. These solutions were prepared with a pH 3 buffer composed of 90 mM acetic acid, 10 mM L-glutamine, and 4.3 mM OHP-β-CD, compounds that were found not to contribute to the CD response. The response at 295 nm was determined to depend linearly on solute concentration (concentration of (S)-(+)-methadone in µg/mL ) 7.42 * detector response; r ) 0.999; factor for R-(-)-methadone is 7.80). The circular dichroism response as function of enantiomeric composition was not assessed. Analysis of the fractions of Figure 3 by chiral CZE and circular dichroism spectroscopy revealed data that were essentially found to be in agreement (Table 1, Figure 4B). With increased methadone concentration and increased optical purity, the measured circular dichroism response became larger. The response was found to be positive for fractions containing an excess of the S-(+) enantiomer and negative for those with an excess of the R-(-) enantiomer. Analysis of fraction 72 of the data depicted in Figure 3 revealed interesting results that merit some discussion. Using chiral CZE (Figure 3C), this fraction was determined to contain 49.2% of R-MET (expressed as relative area %, Table 1) and an ee value for S-MET of 1.7%, this being close to a racemic mixture. On the basis of these data (small enrichment of (S)-methadone), it was expected that circular-dichroism spectroscopy would provide a slightly positive response. The measured signal, however, was negative (Figure 4B), which leads to the conclusion that this fraction must contain an excess of the R-(-) enantiomer. As the methadone concentration for that fraction is high (about 2.2 mM) and the circular-dichroism response was determined to be small, the excess of (R)-(-)-methadone is (31) Han, S. M.; Atkinson, W. M.; Purdie, N. Anal. Chem. 1984, 56, 28272830.

certainly small as well (on the order of 20 µg/mL (65 µM, corresponding to about 3% of total methadone)). Nevertheless, this excess is not directly seen by chiral CZE with data evaluation based upon relative peak areas. Thus, fractions containing almost equal amounts of the two methadone enantiomers cannot be exactly classified by chiral CZE with data evaluation based upon relative peak areas. On the other hand, proper characterization appears to be possible by circular-dichroism spectroscopy. The data presented in Figure 3 revealed that continuous-flow ITP provides partial separation of the two enantiomers. The processing rate was 18.7 mg per hour of racemic methadone, and the enrichment was about 80%. Reduction of the sample load as well as a change in the ITP conditions via increase of the leader pH to 4.7 did not provide better separation of the two enantiomers. However, operation with interrupted flow during electrophoresis was found to provide almost complete separation of the two enantiomers. Data obtained having a leader pH of 4.5 are presented in Figure 5 and Table 1. For that experiment, the leader did contain 0.2% HPMC, and the sample load was about 0.72 mg of racemic methadone. After establishment of the initial conditions (cf. Experimental Section), a constant current of 25 mA was applied for 2 min, 40 mA for an additional 4 min, followed by operation at a constant 1000 V for another 6 min. The final current was 51 mA (power, 51 W). No power was applied during sample collection. Analysis of the collected fractions (Figure 5A, Table 1) revealed that “pure” R-(-)- methadone was present in fractions 96-100, whereas (S)-(+)-methadone in pure form was collected in fractions 86 and 87. In the context of assessing sample purity by CZE, pure (100% in Table 1, case in which second enantiomer could not be detected) should be regarded as > 99.7%. Although absence of imposed flow did allow separation of the enantiomers, fraction collection with counterflow provided some uneven fluid mixing around the outlet ports. This is clearly seen with the monitored Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1847

Figure 7. Chiral interval-flow ZE data at pH 3 with 5 mM OHP-β-CD as the chiral selector. Panel A shows the enantiomeric composition of the fractions as assessed by chiral CZE, and panel B depicts electropherograms of selected, undiluted fractions that were monitored with the HP instrument. Table 3. Selected Chiral ITP Separation Data of Methadone during Multistage Processinga first stage fraction no.

R-METa (rel area %)

72 74 76 78 80 81 82 83 84 85 86

13.8 9.50 16.5 39.9 53.0 59.4 64.6 71.5 78.8 81.4 81.8

second stage

ee R-MET (%)

ee S-MET (%)

fraction no.

R-METb (rel area %)

72.4 81.0 67.0 20.1

66 67 68 69 70 71 72 73 74 75 76

26.7 27.1 27.7 29.5 31.1 34.4 40.6 51.3 73.8 89.1 91.2

6.00 18.9 29.2 43.1 57.6 62.8 63.5

ee R-MET (%)

ee S-MET (%) 46.6 45.7 44.6 41.0 37.8 31.2 18.8

2.6 47.7 78.1 82.4

a Data presented correspond to those of Figures 8A (first stage) and C (second stage). b Corresponding S-MET data could be calculated by 100% - (R-MET).

“saw-tooth” methadone distribution (Figure 5A) instead of the expected plateau-shaped methadone zone obtained under continuous-flow conditions (Figure 3A). In that operational mode, imposed flow was inactivated during the electrophoretic separation, making this approach an interval rather than a continuous process. Continuous- and Interval-Flow Zone Electrophoresis of Methadone Enantiomers. Continuous-flow ZE separations were executed in a buffer comprising 63 mM acetic acid and 7.0 mM 1848 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

(pH 3, κ ) 0.35 mS/cm) and having 5 mM OHP-βCD as chiral selector. Typical data for the enantiomeric composition of the fractions as assessed by chiral CZE are presented in panel A of Figure 6, whereas panel B depicts electropherograms of selected fractions. In that run, the residence time was 5 min, the voltage was kept constant at 900 V, and equilibrium was attained after 25 min of power application. From that time point on, the current was 86 mA (applied electric power, 77.4 W), and

L-glutamine

fraction collection was executed during about 20 min. The rate of counterflow of buffer (line 9 of Figure 2) used was about 1.2-fold higher compared with the total flow of 4 mL/min through the cell. Thus, at the collection outline ports, the chamber fluid was continuously diluted with buffer. The sample was composed of 7.5 mg/mL of racemic methadone dissolved in 80% separation buffer, and the infusion rate was about 1.87 mL/h (throughput: about 14 mg/h). The data presented in Figure 6 and Table 2 reveal that partial separation of the two enantiomers is obtained. Fractions 53-74 were found to be enriched with (S)-(+)methadone, whereas an excess of (R)-(-)-methadone was determined to be in fractions 75-94. Maximum ee values for (S)-(+)methadone and (R)-(-)-methadone were determined to be 67.9% (fraction 54) and 82.0% (fraction 94), respectively. Thus, not surprisingly, enantiomer enrichment was highest at the edges of the overall methadone distribution. Compared to the data obtained by continuous-flow ITP (Figure 3), sample throughput was similar. However, methadone concentration in the collected fractions outside the edges (about fractions 60-85) was four to five times lower (about 0.5 mM). Running conditions with a smaller sample concentration and higher infusion rate were also investigated, efforts that lead to almost pure fractions at the edges of the methadone distribution. However, the methadone concentration in these fractions was not larger than about 40 µg/mL (130 µM) of (R)-(-)-methadone (data not shown). Furthermore, use of small amounts of HPMC in the buffer was found not to improve enantiomer separation. As in the case with ITP, operation with interrupted flow during electrophoresis was also found to provide almost complete separation of the two enantiomers. Typical data obtained are presented in Figure 7 and Table 2. In that experiment a buffer composed of 90 mM acetic acid and 10 mM L-glutamine (pH 2.8, κ ) 0.424 mS/cm) containing 5 mM of the chiral selector and 0.2% HPMC (w/v) was employed. The sample load was about 1.8 mg of racemic methadone. After establishment of the initial conditions (cf. Experimental Section), a constant voltage of 1190 V (current, about 79 mA; power, about 94 W) was applied for 9 min. No power was applied during sample collection. Analysis of the collected fractions revealed that pure (R)-(-)-methadone was present in fractions 92-96, whereas (S)-(+)-methadone in pure form was collected in fractions 70-81. Furthermore, fraction collection in the presence of counterflow (rate about half of the buffer flow rate within the chamber) provided, again, some undesired fluid mixing around the outlet ports. Multistage Continuous-Flow Isotachophoretic Purification of (R)-(-) Methadone. Using continuous-flow instrumentation, the purification of (R)-(-)-methadone by double-stage continuousflow ITP was assessed. First, racemic methadone was processed as described above (Figure 3), with the exception that the run was performed at a constant 700 V (current at equilibrium, 38 mA; power, 26.6 W). Under these conditions, methadone migrated further toward the cathode and was mainly collected in fractions 70-85 (Figure 8A). Enantiomeric separation (Table 3) was found to be comparable to that obtained at 600 V (Figure 3, Table 1). (R)-(-)-methadone was determined to be enriched in fractions 80-86, whereas partial purification of (S)-(+)-methadone was noted in fractions 68-79. Thus, for further processing, fractions 81-85 were combined, thereby producing a sample characterized

Figure 8. Multistage chiral continuous-flow ITP data of (A) first stage with processing of 18.7 mg/h racemic methadone and (B,C) second stage with reprocessing of fractions 81-85 of first stage. Data were assessed and processed as for Figure 3.

by an R-MET value of 70.4% (ee R-MET value of 40.7%) and a methadone concentration of about 2.2 mM. This sample was again infused under continuous-flow ITP conditions (sample flow rate, 9.5 mL/h), processed at a constant 600 V (current, 36 mA; power, 21.6 W), and collected in fractions 66-76. Analysis of the fractions revealed proper ITP behavior of methadone with fraction 75 reaching the ITP plateau concentration (Figure 8B). The latter aspect is documented with the absorbance data of that fraction compared with the corresponding data presented in Figure 3A. Furthermore, CZE analysis of the fractions revealed high enrichment of (R)-(-)-methadone in fractions 74-76 (Figure 8C, Table 3). Without loss of total methadone concentration, fraction 75 was found to be characterized by an R-MET value of 89.1%, this representing a significant purification increase for the second ITP processing stage. The enantiomeric excess ee almost increased Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1849

2-fold to a value of 78.1%. Infusion of fraction 75 into a third continuous-flow ITP system would provide an even higher purity. Thus, using instrumentation comprising multiple stages, complete purification of (R)-(-)-methadone could also be obtained under continuous-flow conditions. CONCLUSIONS In the free-fluid electrophoretic approaches discussed the chiral selector is used as buffer additive. This is distinctly different from separations of stereoisomers based upon liquid chromatography in which β-CD-bonded media are employed.32 OHP-β-CD was used as chiral selector. This compound, covalently bound to silica, has also been reported as being suitable for both preparative4 and analytical-scale (for an example demonstrating the determination of methadone enantiomers in biological fluids refer to ref 33) liquid chromatographic separations of enantiomers. Continuous-flow ZE and ITP using OHP-β-CD as the chiral selector and having electrophoretic mobility differences in the order of 1 × 10-5 cm2/ (V s) are shown to permit partial separation of enantiomers. With the Octopus apparatus and a processing rate of 10-20 mg of racemic methadone per h, enantiomeric enrichment of about 80% is obtained with both methods. Higher purity and no loss of methadone concentration in the product stream is obtained via multistage ITP processing during which selected fractions of one stage are combined and infused as sample into the next operational stage. Furthermore, in an operational mode in which the imposed buffer flow is inactivated during the electrophoretic separation (interval-flow electrophoresis), complete separation and purification of the two enantiomers with a mg/h throughput is (32) Armstrong, D. W.; Ward, T. J.; Armstrong, R. D.; Beesley, T. E. Science (Washington, D.C.) 1986, 232, 1132-1135. (33) Pham-Huy, C.; Chikhi-Chorfi, N.; Galons, H.; Sadeg, N.; Laqueille, X.; Aymard, N.; Massicot, F.; Warnet, J.-M.; Claude, J.-R. J. Chromatogr., B 1997, 700, 155-167. (34) Strickler, A.; Sacks, T. Ann. N. Y. Acad. Sci. 1973, 209, 497-514. (35) Hannig, K.; Wirth, H.; Meyer, B.-H.; Zeiller, K. Hoppe-Seyler’s Z. Physiol. Chem. 1975, 356, 1209-1223. (36) Rhodes, P. H.; Snyder, R. S. U.S. Patent 4752 372, June 1988.

1850 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

attained. Discontinuation of the flow, however, makes this approach an interval rather than a continuous process. On the other hand, repetitive operation in an automated format should be possible, leading to a semicontinuous purification process. The data presented illustrate that zone dispersion originating from the imposed buffer flow, a flow which has a parabolic flow profile with strongest transport in the center of the thin fluid film and vanishing flow at the walls,34,35 has a much larger impact on sample resolution than other effects, including electroosmosis, electrophoretic zone dispersion, and electrohydrodynamics. Deteriorating flow effects can effectively be eliminated using the interval-flow mode. Alternatively, sample dispersion due to residence time variations in the Poiseuille profile can be minimized or abolished by applying the sample near the center of the fluid layer or by using chamber walls which move in the flow direction thereby providing the flow without application of a pressure gradient.36 These approaches, however, are more complex from an instrumental point of view than the investigated interval-flow method. Not addressed in this paper are the removal of OHP-β-CD after separation, e.g., by precipitation with an organic solvent, and the reuse of the chiral selector to minimize processing expenses. ACKNOWLEDGMENT The authors gratefully acknowledge the receipt of the enantiomeric standards of methadone from Dr. P. Baumann, University of Lausanne, Lausanne, Switzerland and the loan of a BioFocus 3000 from its manufacturer, Bio-Rad Laboratories, Life Science Group, Hercules, CA. Circular-dichroism measurements were made at the Institute of Chemistry, University of Neuchaˆtel, Neuchaˆtel, Switzerland (BENEFRI program), and the authors acknowledge the kind assistance provided by Prof. K. Bernauer and Mr. U. Scholten. This work was partly sponsored by the Swiss National Science Foundation. Received for review October 28, 1998. Accepted February 21, 1999. AC981178V