Tamsin KellyPhilip DobleMichael Dawson

Centre for Forensic Science,Faculty of Science,University of Technology,Sydney, (UTS),Sydney, Australia

Chiral separation of methadone, 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP) and2-ethyl-5-methyl-3,3-diphenyl-1-pyrroline (EMDP)by capillary electrophoresis using cyclodextrinderivatives

A stereoselective method was developed for the simultaneous determination ofmethadone and its two major metabolites, 2-ethylidene-1,5-dimethyl-3,3-diphenyl-pyrrolidine (EDDP) and 2-ethyl-5-methyl-3,3-diphenyl-1-pyrroline (EMDP) by capillaryelectrophoresis. Five �-cyclodextrin (�CD) background electrolyte (BGE) additiveswere evaluated for resolution efficiency. The conditions for baseline resolution of eachof the three enantiomer pairs was determined to be 1 mM heptakis-(2,6-di-O-methyl)-�-cyclodextrin (DM�CD) in 100 mM phosphate at pH 2.6. This method represents thefirst successful method for the resolution of the six enantiomers associated with themetabolism of methadone. The utilisation of doubly coated capillaries in conjunctionwith �CD derivatives for a faster separation of the methadone-related enantiomers isalso reported. The coated capillaries were prepared using a polycation of poly(diallyl-dimethylammonium chloride) (PDDAC) and a polyanion of dextran sulfate. Baselineresolution of the methadone enantiomers was achieved with a BGE of 8 mM (2-hydro-xy)propyl-�-cyclodextrin (HP�CD) in 100 mM phosphate at pH 2.6. The migration timesfor the stereoselective methadone separation were approximately 4 min, which repre-sented a reduction by a factor of approximately three, compared to that attained usinganalogous conditions with the uncoated capillary.

Keywords: Chiral separation / Cyclodextrin / Methadone DOI 10.1002/elps.200305418

1 Introduction

Racemic methadone (Meth) is administered to heroinusers undergoing methadone maintenance therapy (MMT)in Australia. The enantiomers of Meth possess differentpharmacological effects with respect to receptor affinity,biotransformation and protein binding. For example,(R)-Meth has been observed to have an affinity for the �1

opioid receptor 10 times greater than that of (S)-Meth [1].

The (R)-enantiomer is almost exclusively responsible foranalgesia and withdrawal management. The main meta-bolic pathways of Meth (Fig. 1) involve N-demethylationfollowed by spontaneous cyclisation. The two main meta-bolites of Meth are 2-ethylidene-1,5-dimethyl-3,3-di-phenylpyrrolidine (EDDP) and 2-ethyl-5-methyl-3,3-di-phenyl-1-pyrroline (EMDP); both of these compoundsare chiral. In MMT patients, 20–60% of the Meth dose isexcreted in urine in 24 h, with up to 33% as unchangeddrug, up to 43% as EDDP and 5–10% as EMDP [2]. Avariety of HPLC [3–5] and CE [2, 6–10] methods have beenpublished detailing the enantioseparation of Meth and/orits major metabolite (EDDP). However, to date no paper

Figure 1. Main metabolic pathways of Meth (asteriskdenotes chiral centre).

Correspondence: Dr. Philip Doble, Centre for Forensic Science,Faculty of Science, University of Technology, Sydney (UTS), POBox 123, Broadway N.S.W. Australia 2007E-mail: philip.doble@uts.edu.auFax: +61-2-9514-1460

Abbreviations: DM�CD, heptakis-(2,6-di-O-methyl)-�-cyclo-dextrin; DS, dextran sulfate; EDDP, 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine; EMDP, 2-ethyl-5-methyl-3,3-diphenyl-1-pyrroline; HP�CD, (2-hydroxy)propyl-�-cyclodextrin; M�CD,methylated �-cyclodextrin; Meth, methadone; MMT, methadonemaintenance therapy; PDDAC, poly(diallyldimethylammoniumchloride); PRs, product resolution; TM�CD, heptakis-(2,3,6-tri-O-methyl)-�-cyclodextrin

2106 Electrophoresis 2003, 24, 2106–2110

2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0173-0835/03/12-1306–2106 $17.50�.50/0

Electrophoresis 2003, 24, 2106–2110 Chiral analysis of methadone and its metabolites by CE 2107

has been published that details the simultaneous stereo-selective analysis of Meth and its two major metabolites(EDDP and EMDP). A stereoselective method capable ofquantifying Meth and its two major metabolites in biolog-ical samples would be of benefit for the monitoring ofMMT patients.

CE with chiral additives is frequently used for the resolu-tion of enantiomeric compounds. A wide variety of chiraladditives are available, such as macrocyclic antibiotics,linear polysaccharides, chiral micelles, ergot alkaloids,crown ethers, and cyclodextrins (CDs). A thorough reviewof chiral separations utilising these compounds in con-junction with CE was performed by Vespalec and Bocek[11]. The authors identified CDs as the most popular chiralselector, owing to the high-separation efficiency andselectivity they generally provide.

CD additives have frequently been used in the chiral separa-tion of enantiomers for forensic toxicology purposes. Ram-seier et. al. [9] developed a chiral CE method for the urinaryanalysis of several amphetamine-related compounds, in-cluding amphetamine, methamphetamine, and 3,4-methyl-enedioxymethamphetamine (MDMA), in the presence ofMeth and its major metabolite, EDDP. The method utilised(2-hydroxy)propyl-�-CD (HP�CD) as the chiral additive ata concentration of 8.3 mM in 75 mM phosphate buffer atpH 2.5. The (R)-enantiomer of Meth was observed tomigrate before its corresponding (S)-enantiomer. Due tothe unavailability of pure enantiomeric standards forEDDP, the authors speculated that (R)-EDDP wouldmigrate ahead of the corresponding (S)-enantiomer [9].

Lanz and Thormann [2] detailed a CZE method for thechiral determination of Meth and EDDP in urine, using aBGE consisting of 4.3 mM HP�CD in 100 mM phosphatebuffer at pH 3. Complete baseline resolution of the Methenantiomers, and partial resolution of the EDDP enan-tiomers were achieved. Baseline chiral separation ofMeth and EDDP in various biological matrices, such asserum, urine and hair, has also been reported using aBGE consisting of 2 mM heptakis-(2,6-di-O-methyl)-�-CD(DM�CD) in 100 mM phosphate buffer at pH 2.3 with 10%methanol [10].

Recent developments in the preparation of coated capil-laries that result in pH-independent EOF allow for thepossibility of very fast cationic separations. The coatingprocedure generally involves an initial coating of a poly-cation, followed by a polyanion. Zakaria et. al. [12] utiliseda doubly coated fused-silica capillary with �CD as apseudostationary phase for the separation of a seriesof 14 aromatic bases by electrokinetic chromatography,with an analysis time of approximately 2 min. The authorsconcluded that the capillaries coated with a polycation ofpoly(diallyldimethylammonium chloride) (PDDAC) and a

polyanion of dextran sulfate (DS) produced stable EOFvalues over the pH range of 3.5–7.5 and �CD concentra-tions from 0 to 10 mM.

Dynamically coated capillaries have also been utilised inthe development of a method for the analysis of the fivemajor alkaloids (morphine, codeine, thebaine, noscapine,and papaverine) present in opium [13]. Two proprietaryreagents were used for the polycation and polyanion,known as CElixir Reagent A and CElixir Reagent B,respectively. Each reagent was consecutively applied tothe uncoated fused-silica capillary, with subsequentseparations achieved using a mixture of HP�CD andDM�CD, at pH 2.5. While CD additives have been usedin association with doubly coated capillaries, previousliterature has focused on achiral applications with theobjective of obtaining a rapid separation method. Fastchiral applications based on CD additives with doublycoated capillaries have not been reported. The primaryobjective of this study was to evaluate the application ofCE using �CD derivatives as chiral additives for the simul-taneous chiral separation of Meth and its two major meta-bolites (EDDP and EMDP). Furthermore, the applicationof doubly coated capillaries was also investigated withthe desire for a fast and robust separation.

2 Materials and methods

2.1 Reagents

All buffers were prepared in Milli-Q grade water by dilutionof stock solutions of phosphoric acid (H3PO4) and sodiumdihydrogen phosphate (NaH2PO4) (Sigma-Aldrich, Syd-ney, Australia). All BGEs were degassed by sonicationunder vacuum and filtered using 0.45 �m nylon syringefilters (Bonnett Equipment, Sydney, Australia) prior touse. rac-Meth, rac-EDDP, rac-EMDP were obtained fromDiagnostic Consultants (Sydney, Australia). Due to theunavailability of the individual enantiomeric standards,peak allocation for each enantiomer could not be per-formed during the method development. Five CD BGEadditives were chosen as candidates for this study. �CD(1135 g/mol), DM�CD (1331 g/mol) and heptakis-(2,3,6-tri-O-methyl)-�-CD (TM�CD, 1430 g/mol) were purchasedfrom Sigma-Aldrich. Methylated �CD (M�CD, CavasolW7 M Pharma, 1310 g/mol) and HP�CD (Cavasol W7 HPPharma, 1400 g/mol) were kindly donated by Wacker-Chemie GmbH (Burghausen, Germany).

2.2 CE

Experiments were conducted on an Agilent ChemstationCapillary Electrophoresis System (Agilent Technologies,Palo Alto, CA, USA), equipped with a diode array detec-

CE

and

CE

C

2108 T. Kelly et al. Electrophoresis 2003, 24, 2106–2110

tor. Uncoated 50 �m ID fused-silica capillaries of totallength of 65 cm and effective length of 56.5 cm were ob-tained from Polymicro Technologies (Phoenix, AZ, USA).Unless stated otherwise, all separations were performedat ambient laboratory temperature of 25�C, with an ap-plied voltage of 30 kV (positive polarity) and detectionwas performed at 195 nm. Aqueous standards of rac-Meth, rac-EDDP, rac-EMDP were prepared in Milli-Q waterto give a concentration of 20 �g/mL. Hydrodynamic injec-tions were used at 50 mbar for 2 s. At the start of theworking day, the capillary was flushed with 0.1 N NaOHfor 5 min, Milli-Q water for 5 min, and then the runningbuffer for 15 min. Between different BGEs, the capillarywas flushed for 15 min with the new BGE. Furthermore, aflush of the running buffer was performed for 3 min beforeeach injection. The coated capillaries were prepared bya method previously described by Zakaria et al [12]. Insummary, the capillary was flushed with 1 N NaOH for30 min, then with Milli-Q water for 15 min before standingfor 30 min. A solution of 1% v/v PDDAC was then flushedthrough the capillary for 15 min, which was allowed tostand for 15 min before a 2 min flush with Milli-Q water.Finally, 1.5% w/v DS solution was flushed through thecapillary for 15 min. After 30 min standing, the capillarywas flushed with Milli-Q water for 5 min. Between differ-ent BGEs, the capillary was flushed for 15 min, and thenallowed to stand for a further 15 min. The position of theEOF was verified for each BGE by injection of an acetonesolution (10% acetone in 10% buffer solution). Unlessstated otherwise, all other experimental conditions wereas described for the uncoated capillaries. The resolution(Rs) of each enantiomer pair was calculated for eachexperimental condition using the migration time for eachpeak and their corresponding baseline peak widths. Incases where partial peak resolution was observed, thebaseline peak widths were estimated by twice the peakwidth from the leading (peak 1) or tailing (peak 2) edge tothe apex of the peak. The product resolution (PRs) valuefor each separation was calculated by Eq. (1).

PRs = Meth Rs�EDDP Rs�EMDP Rs (1)

where Meth Rs, EDDP Rs and EMDP Rs are the Rs valuesfor the Meth, EDDP and EMDP enantiomer pairs, respec-tively.

3 Results and discussion

Preliminary evaluations of each of the five �CD deriva-tives were performed using a 100 mM phosphate bufferat pH 2.5. The concentration of each CD additive was8 mM. The resolution results obtained for each CD deriva-tive are detailed in Fig. 2. The three CDs that gave thehighest resolutions were selected for further development

Figure 2. Rs versus type of �-CD derivative. Conditions:50 �m ID�65 cm uncoated capillary at �30 kV; 100 mM

phosphate, pH 2.5; temperature, 25�C; injection, 2 s at50 mbar; detection, � = 195 nm.

of the method, i.e., DM�CD, HP�CD, and M�CD. Each ofthe selected CDs was tested in concentrations rangingfrom 4 to 25 mM in 100 mM phosphate buffer at pH 2.6and 3.1. These pH values were selected based on pre-viously published papers relating to the enantiosepara-tion of Meth and EDDP enantiomers [2, 9].

HP�CD was found to possess considerable enantio-selectivity for the Meth enantiomers, with Rs values rang-ing from 5.9 to 12.2. However, only partial resolutions(Rs � 1.26) were achievable for the EDDP enantiomerpair and no enantioselectivity (Rs = 0) was observed forEMDP. In general, the enantioselectivity of HP�CD forboth Meth and EDDP increased slightly with decreasingconcentration of HP�CD. Wren and Rowe [14] have stat-ed the greater the affinity that the enantiomers possessfor the CD selector, the lower the CD concentrationrequired for resolution of the enantiomers. Therefore, thissuggests that both Meth and EDDP have a high affinityfor HP�CD. M�CD was found to possess some enantio-selectivity for Meth, with Rs values ranging from 1.26 to3.19. M�CD was capable of discriminating between theEMDP enantiomers to a similar extent, with Rs valuesranging from 1.15 to 2.61. However, no enantioselectivitywas observed for EDDP. A similar trend with respect toCD concentration and enantioselectivity was observedfor M�CD as with HP�CD, i.e., the Rs values increasedwith decreasing M�CD concentration.

DM�CD was enantioselective for all of the enantiomericpairs under all conditions with the exception of 25 mM

DM�CD, where no enantioselectivity for any of the ana-lytes was observed at either pH. In general, the enantio-selectivity of each enantiomer pair increased with decreas-ing DM�CD concentration. Such a pronounced effect wasobserved that the concentration range of DM�CD was

Electrophoresis 2003, 24, 2106–2110 Chiral analysis of methadone and its metabolites by CE 2109

Figure 3. Influence of DM�CD concentration on PRs

using 100 mM phosphate, pH 2.6. Other conditions asin Fig. 1.

extended to include 0.25 mM, 0.5 and 1 mM DM�CD foreach pH value. Similar to the previous two �CDs, theenantioselectivity was greatest for the Meth enantiomers,with Rs values ranging from 1.1 to 5.35 (0.25–15 mM

DM�CD, 100 mM phosphate, pH 2.6 and 3.1), followedby EMDP (Rs values of 0.66–3.7), and then EDDP (Rs

values 0.35–1.97). The effect of DM�CD concentrationon the product resolution (PRs) for the three enantiomericpairs at pH 2.6 is given in Fig. 3. A similar trend wasobserved for the PRs at pH 3.1. Baseline separation of allenantiomeric pairs was achieved with 1 mM DM�CD in100 mM phosphate at pH 2.6. An example electrophero-gram obtained for an aqueous mixture of rac-Meth, rac-EDDP and rac-EMDP (each at 20 �g/mL) using these con-ditions is illustrated in Fig. 4.

For each of the three CD derivatives, the Meth enantio-mers were generally observed to have the greatest reso-lution compared to that of both EDDP and EMDP. Ram-seier et. al. [9] proposed that larger residues positionednear the chiral centre of the Meth molecule facilitatedthe formation of stronger inclusion complexes, and thusincreased resolution of the enantiomers of Meth com-pared to a series of amphetamine analogues. The cyclisa-tion of the compound which occurs during the formationof EDDP and EMDP may therefore influence the formationof an inclusion complex of these analytes with each of theCD derivatives.

The selected CDs were also evaluated using a coatedcapillary, using 100 mM phosphate buffer, pH 2.6 and 3.1.The resolution of the enantiomers under analogous con-ditions to that of the uncoated capillary was dramaticallyreduced. Indeed, Meth was the only pair of enantiomersthat was successfully resolved. The range of Rs valuesobtained for the Meth enantiomers using 4–25 mM

HP�CD in 100 mM phosphate at pH 2.6 and 3.1 for thecoated capillary was 0–5.96. This compares with Rs

values of 5.9–12.2 for the uncoated capillary. The experi-mental conditions that gave the greatest resolution ofthe Meth enantiomers using the coated capillary 8 mM

Figure 4. Example electropherogram for aqueous mixedstandard of rac-Meth, rac-EDDP and rac-EMDP (each at20 �g/mL) with 1 mM DM�CD in 100 mM phosphate,pH 2.6. Migration order: 1, EDDP1; 2, EDDP2; 3, EMDP1;4, EMDP2; 5, Meth1; 6, Meth2. Other conditions as inFig. 1.

HP�CD in 100 mM phosphate, pH 2.6; an exampleelectropherogram for an injection of 20 �g/mL aqueousstandard of rac-Meth is illustrated in Fig. 5. The use ofthe coated capillary permitted a considerable reductionin the analysis time for the Meth enantiomers, with migra-tion times dropping from approximately 13–14 min to4 min when using 8 mM HP�CD in 100 mM phosphate,pH 2.6. Thus, these conditions provide a means by whichchiral analysis of Meth can be rapidly performed.

It is important to note that in contradiction to previousreports [12], the EOF of the doubly coated PDDAC/DSfused-silica capillary was not reproducible, with coeffi-cients of variation ranging between 21 and 25% for asequence consisting of approximately 60–70 injectionsusing buffers of 100 mM phosphate, pH 2.6 and 3.1. Thismay have been due to the DS coating washing off the sur-face of the capillary, resulting in a decrease in the EOFover the duration of the lengthy sequences performed inthis study. Periodic flushing of the capillary with 1.5% DS

2110 T. Kelly et al. Electrophoresis 2003, 24, 2106–2110

Figure 5. Example electropherogram for a 20 �g/mLaqueous standard of rac-Meth with 8 mM HP�CD in100 mM phosphate, pH 2.6. Migration order: 1, Meth1;2, Meth2. Other conditions as in Fig. 1.

solution after the initial coating procedure was found toregenerate the EOF to some extent; however the regen-eration was found to be variable with differences in theEOF observed following each respective flush. One alter-native may be to include a small amount of DS in the run-ning buffer; as used by Lurie et. al. [13] for a coated capil-lary based on CElixir reagents and a BGE consisting of75 mM phosphate, pH 2.5. The investigation of alternativecoating procedures may also prove to be beneficial,including consideration of alternative polycations/poly-anions (Polybrene and poly(vinylsulfonate) [15], CElixirreagents), and the pH of coating solutions (for example,75 mM phosphate buffer, pH 2.5 [13]).

The developed method has potential applications withrespect to the quantitation of Meth related enantiomersin biological matrices, including plasma and hair. Al-though the CE-UV technique may be associated withsensitivities issues when applied to authentic biologicalsamples, the adaptation of this method for CE-MS tech-niques may help to overcome such issues.

4 Concluding remarks

Following the evaluation of several �CD derivatives usingan uncoated fused-silica capillary, a capillary electropho-retic method for the simultaneous chiral analysis of Meth

and its two major metabolites (EDDP and EMDP) wassuccessfully developed. Conditions for the greatest reso-lution of each of the three enantiomeric pairs were deter-mined to be 1 mM DM�CD in 100 mM phosphate buffer,pH 2.6 (each Rs value � 1.8). This method represents thefirst technique for the simultaneous separation of the sixenantiomers related to Meth. This method is of signifi-cance for the monitoring of patients undergoing MMT,and to determine a correlation or otherwise betweenMMT failures and interindividual variation in the levelsof each enantiomer. While the application of doublycoated capillaries in conjunction with �CD derivativeswas not found to be highly successful, a method for therapid stereoselective determination of Meth has beenachieved, with migration times of approximately 4 min.The preliminary results associated with the coated capil-laries suggests that further studies are required in orderto achieve a highly rapid and robust method for thesimultaneous chiral separation of the six enantiomers ofinterest.

Received January 15, 2003

5 References

[1] Kristensen, K., Blemmer, T., Angelo, H. R., Christrup, L. L.,Drenck, N. E., Rasmussen, S. N., Sjøgren, P., Ther. DrugMonit. 1996, 18, 221–227.

[2] Lanz, M., Thormann, W., Electrophoresis 1996, 17, 1945–1949.

[3] 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, 155–163.

[4] Foster, D. J. R., Somogyi, A. A., Bochner, F., J. Chromatogr.B 2000, 744, 165–176.

[5] Kintz, P., Eser, H. P., Tracqui, A., Moeller, M., Cirimele, V.,Mangin, P., J. Forens. Sci. 1997, 42, 291–295.

[6] Lanz, M., Caslavska, J., Thormann, W., Electrophoresis1998, 19, 1081–1090.

[7] Hoffmann, P., Wagner, H., Weber, G., Lanz, M., Caslavska,J., Thormann, W., Anal. Chem. 1999, 71, 1840–1850.

[8] Wind, M., Hoffman, P., Wagner, H., Thormann, W., J. Chro-matogr. A 2000, 895, 51–65.

[9] Ramseier, A., Caslavska, J., Thormann, W., Electrophoresis1999, 20, 2726–2738.

[10] Frost, M., Köhler, H., Blaschke, G., Electrophoresis 1997,18, 1026–1034.

[11] Vespalec, R., Bocek, P., Electrophoresis 1999, 20, 2579–2591.

[12] Zakaria, P., Macka, M., Haddad, P. R., Electrophoresis 2002,23, 1844–1852.

[13] Lurie, I. S., Panicker, S., Hays, P. A., Garcia, A. D., Geer, B. L.,J. Chromatogr. A 2003, 984, 109–120.

[14] Wren, S. A. C., Rowe, R. C., J. Chromatogr. 1992, 603,235–241.

[15] Katayama, H., Ishihama, Y., Asakawa, N., Anal. Chem. 1998,70, 2254–2260.