Separation and online preconcentration by multistep stacking with large-volume injection of anabolic steroids by capillary electrokinetic chromatography using charged cyclodextrins and UV-absorption detection

J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Pawel L. U ban Ca men Ga ca-Ruiz Mangeles Ga ca Ma. Luisa Ma ina Depa amen o de Qumica Anal ica, Facul ad de Qumica, Uni e sidad de Alcal, Alcal de Hena es, Mad id, Spain Sepa a ion and online p econcen a ion by mul is ep s acking wi h la ge- olume injec ion o anabolic s e oids by capilla y elec okine ic ch oma og aphy using cha ged cyclodex ins and UV-abso p ion de ec ion The sepa a ion o h ee common anabolic s e oids (me hyl es os e one, me hand o- s enolone and es os e one) was pe o med o he i s ime by capilla y EKC. Di e - en cha ged CD de i a i es and bile sal s we e es ed as dispe sed phases in o de o achie e he sepa a ion. A mix u e o 10 mmol/L succinyla ed-b-CD wi h 1 mmol/L b- CD in a 50 mmol/L bo a e bu e (pH 9) enabled he sepa a ion o he h ee anabolic s e oids in less han 9 min. Concen a ion LODs, ob ained o hese compounds wi h low abso p ion o UV ligh , we e ~5610–5 mol/L. The use o online e e se mig a ing sample s acking wi h la ge- olume injec ion ( he e ec i e leng h o he capilla y) ena- bled o imp o e he de ec ion sensi i i y. Sensi i i y enhancemen ac o s (SEFs) an- ging om 95 ( o es os e one) o 149 ( o me hyl es os e one) we e achie ed by sin- gle s acking p econcen a ion. Then, he possibili ies o mul is ep s acking o imp o e he sensi i i y o hese analy es we e in es iga ed. SEFs ob ained by double s acking p econcen a ion anged om 138 o 185, enabling concen a ion LODs o 2.79610–7 mol/L ( o me hyl es os e one), 3.47610–7 mol/L ( o es os e one) and 3.56610–7 mol/L ( o me hand os enolone). Al hough online iple s acking p econcen a ion was achie ed, i s epea abili y was e y poo and SEFs o he s udied analy es we e no calcula ed. Key Wo ds: Anabolic s e oids; Capilla y elec okine ic ch oma og aphy; Online mul is ep s acking; Recei ed: Ma ch 10, 2005; e ised: Ap il 13, 2005; accep ed:May 11, 2005 DOI 10.1002/jssc.200500113 2200 U ban, Ga ca-Ruiz, Ga ca, Ma ina 1 In oduc ion Al hough CE shows impo an ad an ages (small quan i y o injec ed sample, high speed and esolu ion and low expendi u e o chemicals) o he analysis o many com- pounds in a g ea a ie y o samples [1], he concen a ion LODs ob ained wi h UV-abso p ion de ec ion a e s ill, in many cases, unsa is ac o y. They can be imp o ed by hyphena ion o CE wi h mo e sensi i e de ec o s, such as LIF [2] o elec ochemical de ec ion [3]. Howe e , he cheapes and he mos popula one is he UV-abso p ion de ec o . Al hough he sho diame e o he capilla y (25– 100 lm) does no allow o measu e abso bance o some compounds which do no possess good ch omopho es, he use o special de ec ion windows (e.g. bubble cells, ze a cells) [4, 5] may sligh ly imp o e sensi i i y wi h op i- cal de ec ion. In addi ion, CE o e s a g ea possibili y o online sample p econcen a ion and enables he au oma- isa ion o he p ocess, which always is desi able in analy- ical chemis y, as well as in o he ields. Online sample concen a ion ep esen s an e ec i e and e sa ile way o enhance concen a ion sensi i i y in CE [6]. The high elec ic ield and unable elec opho e ic mobili y o he analy es can be used o induce elec oki- ne ic ocussing wi hin la ge-injec ion olumes o sample di ec ly on-capilla y p io o de ec ion. Online ocussing is no mally pe o med by selec ing di e en bu e p ope ies o modi y analy e eloci y in wo o mo e sec ions in he capilla y, such as sample and BGE. The e a e ou majo modes o online sample concen a- ion in CE [6]: sample s acking, ansien ITP, sweeping and dynamic pH junc ion. Each me hod elies on a dis inc Co espondence: P o esso Ma. Luisa Ma ina, Depa amen o de Qumica Anal ica, Facul ad de Qumica, Uni e sidad de Al- cal, C a. Mad id Ba celona Km. 33.600, 28871 Alcal de He- na es, Mad id, Spain. Fax: +34-91-8854971. E-mail: [email p o ec ed]. Abb e ia ions: CE-b-CD (DS ~ 3), ca boxye hyla ed-b-CD (de- g ee o subs i u ion ~ 3); CM-b-CD, ca boxyme hyla ed-b-CD; MTD,17b-hyd oxy-17a-me hyl-and os a-1,4-dien-3-one (me- hand os enolone); MTS,17b-hyd oxy-17a-me hyl-4-and os en- 3-one (me hyl es os e one); SEFs, sensi i i y enhancemen ac- o s; Suc-b-CD, succinyla ed-b-CD; TST,17b-hyd oxy-4-and os- en-3-one ( es os e one) EKC Sepa a ion/on-line p econcen a ion by mul i-s ep s acking o s e oids 2201 ocussing mechanism based on di e en elec oly e p op- e ies be ween sample and BGE zones, such as conduc- i i y (ionic s eng h), elec oly e coion mobili y, addi i e concen a ion (analy e–addi i e in e ac ions) and bu e pH, espec i ely. Applica ion o one o hese me hods e- quen ly o e s good sensi i i y enhancemen , a imes, wi hin se e al o de s o magni ude. Sample s acking (in no mal and e e sed pola i y modes, wi h e e se mig a ing micelles wi hou and wi h wa e plug, and head column ield-enhanced sample injec ion wi hou and wi h e e se mig a ing micelles) occu s as ions ac oss a bounda y ha sepa a es egions o he high elec ic ield sample zone and he low elec ic ield back- g ound solu ion zone [7]. Some imes sensi i i y enhancemen ob ained wi h single s acking p ocedu e is limi ed and he e is a need o ob ain lowe concen a ion LODs. The e a e some s udies desc ibing success ul applica ion o double s acking p o- cedu e [8–11], showing ha mul is ep s acking can be an in e es ing al e na i e o be used as online p econcen a- ion me hod. Hence, he possibili ies o mul is ep s acking and he lexibili y o his online p econcen a ion me hod o adjus he desi ed concen a ion LODs, depending on he numbe o s acking s eps and making a comp omise be ween eco e y and pe o mance ( esolu ion and anal- ysis ime), should be in es iga ed. Me hyl es os e one, me hand os enolone and es os e - one a e popula s e oids (see Fig. 1), audly used in spo s. These anabolic s e oids a e also used as medi- cines in he ea men o anaemia, enal insu iciency, endome iosis, he edi a y angioedema and inope able b eas cance [12], which e eals a huge necessi y o de elop new analy ical me hods o hei quan i ica ion. Nowadays, hese anabolic s e oids a e mainly analysed by HPLC [13, 14] o GC [14, 15]. Wi h GC-MS, e y low LODs can be achie ed. Howe e , since some anabolic s e oids and hei me aboli es possess hyd oxyl and ca - bonyl g oups in hei s uc u e, de i a isa ion is neces- sa y, as in gene al sample p e ea men is labou ious and ime-consuming [16]. HPLC and CE a e highly p omising me hods o as s e oid sc eening, CE being much less employed han HPLC [17, 18]. CE has been employed o he analysis o di e en s e oids gene ally wi h MS, luo - escence o UV de ec ion [14, 19–21]. Howe e , he in e - es in using he mos widely employed de ec ion sys em in CE (UV abso bance de ec ion) o he analysis o s e oids has p omo ed he use o p econcen a ion echniques p e- ious o hei analysis due o he ela i ely low expec ed sensi i i y (s e oids show low abso p ion in he UV-Vis egion). In ac , o line p econcen a ion echniques as SPE [22, 23] and online p econcen a ion echniques [6, 24–27] ha e been employed. In spi e o he esul s ob ained o he analysis o s e oids by CE, sepa a ion o any h ee anabolic s e oids s udied in his wo k has p e- iously been epo ed. Only he sepa a ion o es os e one and/o me hyl es os e one om o he s e oids has been epo ed [6, 16, 24, 28–30]. Due o he neu al na u e o hese analy es hey need o be sepa a ed by capilla y EKC, which is based on he di e en dis ibu ion o ana- ly es be ween a dispe sed phase (CDs, micelles, e c.) and he mobile phase (bu e solu ion). The aim o his wo k was o op imize he expe imen al condi ions enabling he sepa a ion o he h ee abo e- men ioned anabolic s e oids by EKC, and o de elop a me hod o hei online p econcen a ion in EKC in o de o achie e a sensi i e UV-abso p ion de ec ion o hese low abso ben analy es, his la e pu pose in ol ing he s udy o he possibili ies o mul is ep s acking. 2 Ma e ials and me hods 2.1 Chemicals and samples The s anda ds o 17b-hyd oxy-17a-me hyl-4-and os en- 3-one (me hyl es os e one, MTS), 17b-hyd oxy-17a- me hyl-and os a-1,4-dien-3-one (me hand os enolone, J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 1. S uc u es and names o he h ee anabolic s e - oids s udied in his wo k. O iginal Pape 2202 U ban, Ga ca-Ruiz, Ga ca, Ma ina MTD) and 17b-hyd oxy-4-and os en-3-one ( es os e one, TST) we e pu chased om Fluka (Buchs, Swi ze land) and s o ed a 48C. b-CD and succinyla ed-b-CD (Suc-b- CD (deg ee o subs i u ion (DS) ~3)) we e also supplied by Fluka. Ca boxyme hyla ed-b-CD (CM-b-CD (DS ~3)) and ca boxye hyla ed-b-CD (CE-b-CD (DS ~3)) we e om Cyclolab (Budapes , Hunga y). Sodium bile sal s o cholic acid, au ocholic acid and au odeoxycholic acid we e pu chased om Sigma (S . Louis, MO, USA) and ha o deoxycholic acid om Fluka. Bo ic acid (Fluka), sodium dihyd ogen phospha e (Me ck, Da ms ad , Ge - many) and sodium hyd oxide (Pan eac, Ba celona, Spain) we e used o p epa e he bu e s. HPLC-g adien g ade me hanol was om Pan eac. Dis illed wa e used h oughou his wo k was also pu i ied wi h he Milli-Q sys- em (Millipo e, Bed o d, USA). The 50 mmol/L bo a e bu e (pH 9) was p epa ed by dis- sol ing he app op ia e amoun o bo ic acid in wa e and by adding 1 mol/L sodium hyd oxide o he esul ing solu- ion in o de o ob ain he desi ed pH alue. Then, he sui- able amoun s o CDs we e dissol ed in he bu e solu- ion. The s anda ds o he s udied compounds we e dis- sol ed in me hanol, and hen dilu ed in wa e o ob ain he desi ed concen a ion. Dissolu ion o solids was acili a ed by he use o an ul a- sonic ba h (Raypa, Ba celona, Spain). The pH o each bu e was adjus ed using a 654 pH me e (Me ohm, He - isau, Swi ze land). Bo h, samples and bu e s, we e il- e ed p io o he analysis using sy inge il e s (Ti an Fil a- ion Sys ems; nylon, po e size 0.45 lm, il e size 13 mm). 2.2 Ins umen a ion A CE ins umen (Agilen Technologies, Waldb onn, Ge - many) equipped wi h an oncolumn diode a ay de ec o was used. The con ol o he ins umen and da a acquisi- ion we e achie ed by a PC wi h he 3D-CE ChemS a ion so wa e (A.09.03, Agilen ). Uncoa ed used-silica capilla y was employed: e ec i e capilla y leng h, 50 cm; o al capilla y leng h, 58.5 cm; ID, 75 lm; OD, 375 lm (Sugelabo S.A., Mad id, Spain). The sample and ope a ing ials we e se ed om a 48- ial ca ousel, designed o ypical 1 mL snap polyp opylene ials. 2.3 Analysis p ocedu e Expe imen s wi hou s acking we e pe o med by injec ing he sample (50 mba o 3 s) ollowed by he sepa a ion bu e (50 mba o 3 s). The applied ope a ion scheme used o achie e single, double and iple online s acking is shown in Table 1. The i s s age o each un is capilla y condi ioning wi h 0.1 mol/L NaOH (1 ba o 5 min), Milli-Q wa e (1 ba o 2 min) and unning bu e (1 ba o 5 min). Then, he sample was injec ed o ill 50 cm o he capilla y leng h ( om he inle o he de ec ion window). The injec- ion ime was calcula ed om he ea anged Poiseuille’s equa ion: ¼3200 Ll d2Pð1Þ whe e Pis he injec ion p essu e (mba ), he ime du a- ion o he p essu e (s), d he ID o he capilla y (lm), l he J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Table 1. Analysis s eps in he online mul is ep s acking and sepa a ion o MTS, MTD and TST. Running bu e : 10 mmol/L Suc- b-CD, 1 mmol/L b-CD, 50 mmol/L bo a e bu e a pH 9. Du ing sample concen a ion, –20 kV indica es nega i e pola i y and s1, s2 and s3 a e imes o he successi e concen a ion s eps (176, 209 and 211 s). Used ials: S, sample; B1–B5, unning bu e ; N, 0.1 mol/L NaOH; Q, Milli-Q wa e and W1, W2, was e S eps Inle Ou le Single s acking Double s acking T iple s acking P econdi ioning N Q B1 W1 W1 W2 5 min 0.1 mol/L NaOH 2 min wa e 5 min unning bu e (1 ba ) 5 min 0.1 mol/L NaOH 2 min wa e 5 min unning bu e (1 ba ) 5 min 0.1 mol/L NaOH 2 min wa e 5 min unning bu e (1 ba ) Sample injec ion S B3 50 mba , 332 s (50 cm o capilla y) 50 mba , 332 s (50 cm o capilla y) 50 mba , 332 s (50 cm o capilla y) Bu e injec ion B2 B3 50 mba , 5 s 50 mba , 5 s 50 mba , 5 s Zone concen a ion B2 B3 –20 kV, s1 –20 kV, s1 –20 kV, s1 Sample injec ion S B3 – 50 mba , 332 s 50 mba , 332 s Bu e injec ion B2 B3 – 50 mba , 5 s 50 mba , 5 s Zone concen a ion B2 B3 – –20 kV, s2 –20 kV, s2 Sample injec ion S B3 – – 50 mba , 332 s Bu e injec ion B2 B3 – – 50 mba , 5 s Zone concen a ion B2 B3 – – –20 kV, s3 Sepa a ion B4 B5 20 kV 20 kV 20 kV EKC Sepa a ion/on-line p econcen a ion by mul i-s ep s acking o s e oids 2203 leng h o he sample injec ion plug (mm), g he iscosi y o he bu e (cP) and L he capilla y leng h (cm). The injec- ion ime necessa y o ill he e ec i e leng h o he capil- la y (50 cm) calcula ed by means o Eq. (1) was 332 s. This calcula ed ime was e y simila o he expe imen al ime es ima ed by illing he capilla y wi h a solu ion o a mix u e o he s e oids (~3610–4 mol/L), a e illing he capilla y wi h bu e solu ion, and egis e ing he signal a 245 nm which inc eased when he s e oids achie ed he de ec ion window. A e sample injec ion, a sho zone o he unning bu e was injec ed o p e en loss o he ana- ly e. Then, e e se pola i y ol age (20 kV) was applied. These h ee s eps we e applied once, wice o h ice, depending on he numbe o s acking s eps. Sepa a ion o he mix u e was achie ed in he las s acking s ep, when no mal (posi i e) pola i y ol age (20 kV) was applied. Be o e he i s use, he capilla y was condi ioned by washing i wi h 1 mol/L NaOH (1 ba o abou 1 h) and hen wi h Milli-Q wa e (1 ba o 10 min). Du ing he whole analysis, a ious ials we e used (see Table 1). Di e en bu e ials we e employed o s acking and di e en o sepa a ion (see Table 1) so as o a oid con amina ion o he sepa a ion bu e which could inc ease he back- g ound noise. The bu e ial B1 (used o capilla y p e- condi ioning) was o 2 mL olume since his ial does no equi e equen eplacemen . Two-millili e ials we e also employed o 0.1 mol/L NaOH (N) and wa e s o e (Q) and o he was e deposi ion (W1, W2). One-millili e s anda d plas ic ials we e used o he samples/s an- da ds and bu e solu ions. A empe a u e o 158C was used in his wo k. Mos o he ope a ion s eps we e p o- g ammed using he p econdi ioning able in he ChemS a- ion so wa e. Howe e , he “CE TimeTable” was used o he acquisi ion o he cu en in ensi y cu es du ing he s acking p econcen a ion. Since he s udied compounds sligh ly di e in hei UV- abso p ion maxima, a single wa eleng h was chosen o de ec all o hem (245 nm € 5 nm), and he e e ence alue was swi ched o since i did no gi e any imp o e- men . 2.4 Da a ea men The cu es co esponding o he a ia ion o he cu en in ensi y as a unc ion o he ime o applica ion o e e se pola i y we e ob ained o mix u es o he h ee s e oids in which each componen was a concen a ions anging om 5610–7 o 1610–5 mol/L. Then, he 4 h o de polyno- mial cu es (ax4+bx3+cx2+dx +e) we e adjus ed and he in lexion poin x- alue was calcula ed making he 2nd o de de i a i e om he 4 h o de i ed polynomial cu e (12ax2+6bx +2c) equal o 0. One o he wo solu ions o he 2nd o de de i a i e was conside ed and he ollowing equa ion was used: x¼6bþ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 36b296ac p 24að2Þ whe e a,band ca e he coe icien s a x4,x3and x2in he 4 h o de i ed polynomial cu e. LODs and LOQs in concen a ion (mol/L) o he s e oids in a 1% / me hanol/wa e solu ion we e expe imen ally cal- cula ed acco ding o he ollowing equa ions [31]: LOD ¼3Sa0=b0ð3Þ LOQ ¼10Sa0=b0ð4Þ whe e Sa0is he s anda d e o o he in e cep and b9, he slope o he s aigh line, bo h ob ained by ANOVA. Expe imen al da a we e ea ed using O igin 7.0 (O igin- Lab Co po a ion) and Excel 7.0 (Mic oso ) so wa e. Theo e ical es ima ion o LOD wi h mul is ep s acking was achie ed using Eq. (5): LODn¼LOD0 n ð5Þ whe e LOD0is he LOD ob ained wi hou s acking, n he numbe o s acking s eps (1, 2, 3…) and he s acking ampli ica ion ac o exp essed as: ¼ s 0ð6Þ whe e sis he sample injec ion ime in he single s acking s ep and 0 he injec ion ime in he analysis wi hou s ack- ing. 3 Resul s and discussion 3.1 Sepa a ion condi ions The sepa a ion o a mix u e o neu al MTS, MTD and TST was achie ed a e a se o ials wi h a ious dispe sed phases (CDs o bile sal s) added o bo a e bu e a pH 9, which enabled he e e se mig a ing selec o s o pass h ough he de ec ion window when no mal pola i y is applied. Th ee anionic CDs (CM-b-CD, Suc-b-CD and CE-b-CD) and ou bile sal s (sodium chola e, sodium deoxychola e, sodium au ochola e and sodium au o- deoxychola e) we e in es iga ed. Figu e 2 shows he elec- ophe og ams ob ained o a mix u e o MTS, MTD and TST a a concen a ion o ~3610–4 mol/L in each compo- nen when adding CM-b-CD, Suc-b-CD o CE-b-CD o he elec oly ic solu ion. I can be obse ed ha CM-b-CD and Suc-b-CD a a 10 mmol/L concen a ion enabled he sep- a a ion o he h ee neu al compounds (wi h CM-b-CD, he esolu ion be ween he i s and he second mig a ing peaks (Rs12) was 1.2 and he esolu ion be ween he sec- J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim 2204 U ban, Ga ca-Ruiz, Ga ca, Ma ina ond and he hi d mig a ing peaks (Rs23) was 1.3, whe eas wi h Suc-b-CD, Rs12 = 1.2 and Rs23 = 1.5). The anabolic s e oids passed h ough he de ec o in he ollowing o de : TST, MTS and MTD. Howe e , unde he expe i- men al condi ions ied, CE-b-CD did no enable he sep- a a ion o he mix u e (Rs12 =Rs23 = 0). On he o he hand, om he ou bile sal s es ed, only sodium au o- deoxychola e enabled he sepa a ion o he s udied com- pounds (Rs12 = 1.9 and Rs23 = 2.0) gi ing ise o a di e en mig a ion o de wi h espec o he use o CDs: TST, MTD and MTS, as i is also shown in Fig. 2. The e ec o using mix u es o dispe sed phases was in es iga ed in o de o imp o e he sepa a ion selec i i y ob ained unde he abo e condi ions. Fo his eason, a second CD, he neu al b-CD, was added o he elec oly ic solu ion. Figu e 3 shows ha he p esence o b-CD a a 1 mmol/L concen a ion signi ican ly imp o ed he sepa a- ion o he h ee s e oids. The mechanism h ough which he sepa a ion o he anabolic s e oids was imp o ed by addi ion o a neu al CD is based on he simul aneous in e ac ion o hese compounds wi h he anionic and he neu al dispe sed phases. Thus, hese neu al compounds acqui e mobili y due o hei in e ac ion wi h he anionic dispe sed phase (anionic CD o bile sal ) which, unde he no mal pola i y mode, mig a es owa d he anode. On he o he hand, hei in e ac ion wi h he neu al CD, which mig a es wi h he EOF due o i s elec oneu al na u e, enables hei mig a ion owa ds he ca hode gi ing ise o peaks loca ed in he elec ophe og am be ween he sig- nals co esponding o he EOF and he anionic CDs. Since he in e ac ion o each s e oid wi h he anionic and/o neu- al dispe sed phase is di e en , he disc imina ion inc eases wi h he addi ion o he neu al dispe sed phase imp o ing hei sepa a ion. Thus, he addi ion o b-CD o he CE-b-CD sys em enabled o obse e he sepa a ion o he analy es (Rs12 = 1.3 and Rs23 = 1.0) while he addi ion o b-CD o CM-b-CD, Suc-b-CD o he bile sal ga e ise o an inc ease in he esolu ion. In ac , he use o 1 mmol/L b-CD wi h 25 mmol/L au odeoxychola e enabled o ob ain esolu ions o 2.5 and 2.0 o he i s –second and sec- ond– hi d peaks, espec i ely, al hough he b oades peaks o he analy es we e ob ained unde hese condi- ions (see Fig. 3). Al hough he mix u e o 10 mmol/L CM- b-CD wi h 1 mmol/L b-CD o e ed an excellen sepa a ion o he analy es (Rs12 = 17.0 and Rs23 = 4.1), he mix u e o 10 mmol/L Suc-b-CD and 1 mmol/L b-CD (Rs12 = 7.8 and Rs23 = 4.8) was chosen o u he wo k because i p o- ided simila shape and esolu ion o he peaks o he h ee analy es. Unde hese expe imen al condi ions, con- cen a ion LODs o 4.79610–5 mol/L, 5.17610–5 mol/L and 6.17610–5 mol/L we e ob ained o TST, MTS and MTD espec i ely. In o de o imp o e he UV-abso p ion de ec ion sensi i i y, ob ained o hese h ee s e oids, online sample p econcen a ion by s acking wi h la ge- olume injec ion was applied. J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 2. Sepa a ion o a mix u e o MTS, MTD and TST (~3610–4 mol/L in each componen in 5% / me hanol/wa e ) wi h a 50 mmol/L bo a e bu e a pH 9 con aining an anionic subs i u ed b-CD o he bile sal au odeoxycholic acid. Fused-silica capil- la y, 58.5 cm (50 cm o he de ec ion window), 75 lm ID (375 lm OD); ol age: 20 kV (cu en in ensi y o each sepa a ion bu e is speci ied in he igu e in pa en hesis); injec ion: 50 mba , 3 s and bu e injec ion du ing 3 s; empe a u e: 158C; UV-abso p ion de ec ion a 245 nm (€5 nm). EKC Sepa a ion/on-line p econcen a ion by mul i-s ep s acking o s e oids 2205 3.2 Online p econcen a ion by single s acking Since he selec ed sepa a ion bu e con ained anionic CDs wi h e e se mig a ion in he no mal pola i y mode, he ollowing s eps we e conside ed o pe o m a e e se pola i y s acking mode (see Table 1): (i) sample injec ion o ill he e ec i e leng h o he capilla y, (ii) injec ion o he unning bu e o a oid di usion o he analy es in he ea ly s acking s age, (iii) zone concen a ion applying a ime o e e se pola i y ol age (–20 kV) and (i ) analy e sepa a- ion wi h no mal pola i y ol age (20 kV). The key pa ame e o he p econcen a ion p ocess was obse ed o be he ime o applica ion o he e e se pola - i y ol age (s acking ime). I could be ound in he li e a- u e ha he s acking ime should be ha which allows eaching om 70 o 99% o he cu en in ensi y o he bu - e solu ion [32–36]. The ini ial c i e ion o choose he s acking ime alue was o selec he ime co esponding o he in lexion poin o he cu en in ensi y e sus ime o applica ion o e e se pola i y cu e. This poin was calcu- la ed a e i ing he expe imen al cu en in ensi y da a o a 4 h o de polynomial cu e and making he second de i a i e o he i ed unc ion equal o 0 (see Sec ion 2.4). Al hough i ing he expe imen al cu en in ensi y da a o a 3 d o de polynomial was also achie ed, he i - ing was much be e in he case o he 4 h o de unc ions ( >0.999) being addi ionally less dependen on he ange o he expe imen al cu en in ensi y alues selec ed o be adjus ed. The e o e, he 4 h o de polynomial was used o u he op imisa ion o he s acking ime. In addi ion, due o he s ong dependence o he in lexion poin o he i ed cu e wi h he sample sol en , he use o wa e as sample sol en was desi able du ing he s acking p econ- cen a ion, as i is gene ally known. Howe e , due o he insolubili y o he s udied s e oids in wa e , he use o 1% me hanol was necessa y o solubilise hem be o e he p econcen a ion s age. Unde all hese condi ions, a s acking ime o 150 s was ob ained o a mix u e o MTS, MTD and TST, 10–6 mol/L o each componen . Howe e , di e en imes (calcula ed as men ioned abo e) we e ob ained o di e en concen a ions o he analy es (155 s o 5610–7 mol/L, 150 s o 1610–6 mol/L and 136 s o 5610–6 mol/L). Then, in o de o apply he same s acking ime o mix u es wi h di e en concen a ions o analy es, he a e age o he in lexion poin imes ob ained o mix u es o concen a ions anging om 5610–7 o 5610–6 mol/L, was aken (147 s). This ime co esponded o app oxima ely 75% o he cu en in ensi y o he bu e solu ion. The elec ophe og am co esponding o he injec ion o a mix u e o he h ee anabolic s e oids using his alue as s acking ime enabled o obse e ha when his selec ed s acking ime was applied in a single s ack- ing un, he analy es we e no sepa a ed and pa ially o e lapped wi h he EOF peak ( esul s no shown). Sub- sequen ly, he applica ion o highe s acking imes co e- sponding o pe cen ages o cu en in ensi y highe han 75% was a emp ed. In o de o achie e ha he p e- iously selec ed ime was mul iplied by he ac o s o 1.1 (~80% o he cu en in ensi y o he bu e solu ion), 1.2 (~90%), 1.3 (~95%) and 1.4 (~100%). In all hese cases, he sepa a ion o he h ee peaks co esponding o TST, MTS and MTD could be obse ed enabling o ob ain he calib a ion plo s o he s udied analy es. Since a simila beha iou was obse ed o he h ee analy es, as an example, Fig. 4 shows he calib a ion plo s o TST ( he 1s peak in he elec ophe og am) when mix u es o he J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 3. Sepa a ion o a mix- u e o MTS, MTD and TST (~3610–4 mol/L in each com- ponen in 5% / me hanol/ wa e ) wi h a 50 mmol/L bo a e bu e a pH 9 adding he na i e b-CD o he anionic CDs o he sodium au odeoxychola e. Expe imen al condi ions as in Fig. 1. 2206 U ban, Ga ca-Ruiz, Ga ca, Ma ina h ee s e oids we e injec ed using, as s acking ime, he abo e alue (147 s) mul iplied by he ac o s om 1.1 o 1.4 (s acking imes anged om 162 o 206 s). I can be clea ly seen ha he ac o s 1.1 ( 2= 0.984), 1.2 ( 2= 0.995) and 1.3 ( 2= 0.975) showed good linea i y and simila sensi i i y (slope o he calib a ion lines), which means ha he e is no loss o he analy es when e e se pola i y is applied. Howe e , he use o he long- es s acking ime (206 s) caused losses o he analy es by he inle posi ion. Since he use o he sho es s acking ime (162 s) p oduced b oad peaks, caused by a oo long s acked zone, a ac o o 1.2 ( = 0.997) was chosen o u he expe imen s because he esul an s acking ime, 176 s, p oduced he highes sensi i i y and we ensu ed ha any pa o he analy e zone is no escaping om he capilla y. Using he selec ed s acking ime (176 s, i.e. he ime o each ~90% cu en in ensi y o he bu e solu ion) o online single s acking p econcen a ion, he LOD and LOQ we e calcula ed om he calib a ion lines ob ained plo ing co ec ed peak a eas (a ea/mig a ion ime) e sus concen a ion ( i e alues anging om 5610–7 o 5610–6 mol/L). The co ec ion o he peak a eas was made o compensa e he a ia ions o mig a ion ime o di e en concen a ions o he analy es [37]. Table 2 shows he sensi i i y enhancemen ac o s (SEFs) ob ained o he h ee analy es s udied unde hese condi ions. The lowes SEF was ob ained o he i s peak (TST, SEF 95) whe eas he second and he hi d peaks exhibi ed highe SEF alues (149 o MTS and 139 o MTD), which we e highe han he heo e ical signal ampli ica ion alue cal- cula ed om Eq. (6), which is 111. Howe e , he ob ained SEF alues a e close o 170- old analy e concen a ion achie ed by Chun and Chung [38] when illing he whole capilla y wi h he sample (only he e ec i e leng h o he capilla y was illed wi h he sample in his wo k). 3.3 Online p econcen a ion by mul is ep s acking The heo e ical LODs es ima ed o TST, MTS and MTD, when using mul is ep s acking p econcen a ion, a e shown in Fig. 5. These alues we e ob ained om he LODs o hese compounds wi hou p econcen a ion (see Table 2) and he expec ed ampli ica ion ac o ( om Eq. (6), 332/3 = 111). I can be obse ed ha he e is a signi i- J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 4. Calib a ion cu es ob ained plo ing he peak a ea as a unc ion o TST concen a- ion ( om 5610–7 o 5610–6 mol/L) o a ious ac o s ( om 1.1 o 1.4) employed o mul iply he a e age in lexion poin o he cu en in ensi y cu es. Vol age: 20 kV; injec ion: 50 mba , 332 s and bu e injec ion du ing 5 s; empe a u e: 158C; UV-abso p- ion de ec ion a 245 nm (€5 nm). Table 2. LODs and LOQs exp essed in mol/L, and SEF o LODs when he sepa a ion o he h ee anabolic s e oids was pe - o med wi hou s acking and wi h single and double s acking TST MTS MTD LOD LOQ SEF LOD LOQ SEF LOD LOQ SEF Wi hou s acking n=4 4.79610–5 (€1.12610–5) 1.60610–4 (€0.37610–4) – 5.17610–5 (€1.51610–5) 1.72610–4 (€0.50610–4) – 6.17610–5 (€1.94610–5) 2.06610–4 (€0.65610–4) – Single s acking n=5 5.02610 –7 (€4.84610 –7 ) 1.67610–6 (€1.62610 –6 ) 95 3.47610–7 (€0.81610–7 ) 1.16610 –6 (€0.27610 –6 ) 149 4.42610–7 (€2.60610 –7 ) 1.47610 –6 (€0.87610 –6 ) 139 Double s acking n=5 3.47610 –7 (€1.53610 –7 ) 1.16610–6 (€0.51610 –6 ) 138 2.79610–7 (€1.64610–7 ) 9.30610 –7 (€5.46610 –7 ) 185 3.56610–7 (€1.56610 –7 ) 1.19610 –6 (€0.52610 –6 ) 173 EKC Sepa a ion/on-line p econcen a ion by mul i-s ep s acking o s e oids 2207 can dec ease in he heo e ical LODs o single and dou- ble s acking, while iple and mo e exhaus i e s acking s eps seem o be less e ec i e o imp o ing he sensi i - i y. Table 1 also shows he s eps ollowed o achie e double and iple s acking p econcen a ions. Again, he s acking ime o each s acking s ep was he c ucial pa ame e o be calcula ed o online mul is ep s acking p econcen a- ion. The s acking imes o he second and hi d s acking s eps we e calcula ed by mul iplying he espec i e in lex- ion poin x- alues, ob ained om cu en in ensi y cu es, by he ac o o 1.2 (op imised o he i s s acking s ep). Thus, successi e s acking imes o 176, 209 and 211 s we e calcula ed o single, double and iple s acking, espec i ely. Al hough he s acking ime inc eased wi h he numbe o s acking s eps, i seemed o s abilize wi h he numbe o s acking s eps. LOD and LOQ o he h ee anabolic s e oids when using online double s acking p econcen a ion (see Table 2) we e also calcula ed using co ec ed peak a eas. In his case, SEF alues ob ained (138, 185 and 173 o TST, MTS and MTD, espec i ely) we e lowe han he expec ed heo e ical ampli ica ion o he double s acking, which is 222. This esul could be due o he e o s asso- cia ed wi h he in eg a ion o ailing peaks and he loss o esolu ion which became conside able when using double s acking p econcen a ion (see Fig. 6). The compa ison o he heo e ically es ima ed LOD o TST (Fig. 5) wi h he co esponding expe imen al alue (Table 2), con i med his disc epancy which can be jus i ied conside ing ha Eq. (5) does no ake in o accoun he numbe o heo e i- cal pla es (N) ha usually dec eases wi h he numbe o s acking s eps. In ac , N o he las mig a ing compound (MTD), which co esponds o he peak mo e a ec ed by b oadening e ec s, was 11 000, 11 500 and 7000 o he single, double and iple s acking espec i ely. Quan i ica ion o he iple s acking was no possible because o i s poo epea abili y, especially o he lowes concen a ions. Peak symme y and esolu ion dec eased in his case, as i can be obse ed in Fig. 6, causing huge e o s du ing in eg a ion o peaks. In addi ion, online p e- concen a ion by iple s acking seems o be e y suscep- ible o changes in he expe imen al condi ions. Tempe a- u e a ia ions, especially in he un he mos a ised pa o he capilla y, in luence he iscosi y o bu e and sample a ec ing he analy e eco e y. This makes he iple s ack- ing p econcen a ion a doub ul way o imp o e sensi i i y o he analysis. In addi ion, he analysis wi h mul is ep p e- concen a ion is much longe , almos 1 h in he case o he iple s acking. 4 Concluding ema ks The sepa a ion o he h ee s uc u ally ela ed s e oids, me hyl es os e one, me hand os enolone and es os e - one, has been pe o med o he i s ime by EKC using a dual CD sys em based on an anionic CD and a neu al one (i.e., Suc-b-CD + b-CD) as dispe sed phases. Due o he low abso p ion o UV ligh o hese anabolic s e oids, hey we e used as model compounds o s udy he possibi- li ies o online mul is ep s acking p econcen a ion wi h la ge- olume injec ion in o de o imp o e hei sensi i i y using UV-abso p ion de ec ion. I has been demons a ed J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 5. Theo e ically es i- ma ed LOD wi h mul is ep s acking. LOD o he analysis wi hou s acking co esponds o he alue: 4.79610–5 mol/L. Ampli ica ion o he signal: 111 imes (332/3 s). 2208 U ban, Ga ca-Ruiz, Ga ca, Ma ina ha single and double s acking allow SEFs om 100 o 150 o single s acking and om 140 o 190 o double s acking. In spi e o his, iple s acking was no epea able due o mul iplica ion o e o s associa ed wi h li le changes o expe imen al condi ions. The au ho s hank he Minis e io de Ciencia y Tecnologa (Spain) o p ojec BQU2003-03638. D . C. Ga ca-Ruiz hanks o he same ins i u ion o he con ac om he Ramn y Cajal P og am (RYC-2003-001). P. L. U ban hanks he Eu opean Union o his E asmus ellowship. 5 Re e ences [1] Khaledi, M. G., High Pe o mance Capilla y Elec opho esis: Theo y, Techniques, & Applica ions, John Wiley & Sons, Canada 1998. [2] Kuij , J., Ga ca-Ruiz, C., S oombe g, G. J., Ma ina, M. L., A iese, F., B inkman, U. A. Th., Gooije , C., J. Ch oma og . A2001, 907, 291–299. [3] Woods, L. A., Roddy, T. P., Paxon, T. L., Ewing, A. G., Anal. Chem. 2001, 73, 3687–3690. [4] Go don, G. B., Tella, R. P., Ma ins, H. A. S., Hewle Pack- a d J. 1995, 62–70. [5] Che e , J. P., an Soes , R. E. J., U sem, M., J. Ch oma- og . 1991, 543, 439–449. [6] B i z-McKibbin, P., Ichihashi, T., Tsubo a, K., Chen, D. D. Y., Te abe, S., J. Ch oma og . A 2003, 1013, 65–76. [7] Kim, J.-B., Te abe, S., J. Pha m. Biomed. Anal. 2003, 30, 1625–1643. [8] Plma sd i , S., Edholm, L.-H., J. Ch oma og . A 1995, 693, 131–143. [9] Plma sd i , S., Ma hiasson, L., Jonsson, J. A., Edholm, L.-E., J. Cap. Elec. 1996, 5, 255–260. [10] Plma sd i , S., Tho da son, E., Edholm, L.-E., Jnsson, J. ., Ma hiasson, L., Anal. Chem. 1997, 69, 1732–1737. [11] Plma sd i , S., Ma hiasson, L., Jnsson, J. ., Edholm, L.-E., J. Ch oma og . B 1997, 688, 127–134. [12] Yesalis, C. E. (Ed.), Anabolic S e oids in Spo and Exe cise, Human Kine ics Publishe s, Champaign 1993. [13] Gonzalo-Lumb e as, R., Izquie do-Ho nillos, R., J. Pha m. Biomed. Anal. 2003, 31, 301–308. [14] Appelblad, P., I gum, K., J. Ch oma og . A 2002, 955, 151– 182. [15] T ou , G. J., Kazlauskas, R., Chem. Soc. Re . 2004, 33,1– 13. [16] Amundsen, L. K., Kokkonen, J. T., Ro io, S., Si n, H., J. Ch oma og . A 2004, 1040, 123–131. [17] Smy h, W. F., B ooks, P., Elec opho esis 2004, 25, 1413– 1446. [18] Go og, S., Anal. Sci. 2004, 20, 767–782. [19] Poole, S. K., Poole, C. F., J. Ch oma og . A 1996, 749, 247–255. [20] Shihabi, Z. K., F iedbe g, M. A., Elec opho esis 1997, 18, 1724–1732. [21] Isaaq, H. J., Elec opho esis 1997, 18, 2438–2452. [22] No, S., Bhle , J., Kelle , E., F ahm, A. W., J. Pha maceu . Biomed. Anal. 1998, 18, 471–476. [23] Rao, L. V., Pe e sen, J. R., Bissell, M. G., Oko odudu, A. O., Mohammad, A. A., J. Ch oma og . B 1999, 730, 123–128. [24] Mon on, M. R. N, O suka, K., Te abe, S., J. Ch oma og . A 2003, 985, 435–445. [25] B i z-McKibbin, P., Te abe, S., J. Ch oma og . A 2003, 1000, 917–934. [26] Qui ino, J. P., Te abe, S., J. Ch oma og . A 1998, 798, 251– 257. [27] Qui ino, J. P., Te abe, S., O suka, K., Vincen , J. B., Vigh, G., J. Ch oma og . A 1999, 838, 3–10. J. Sep. Sci. 2005, 28, 2200–2209 www.jss-jou nal.de i2005 WILEY-VCH Ve lag GmbH & Co. KGaA, Weinheim Figu e 6. Compa ison o he elec ophe og ams ob ained by single, double and iple s ack- ing o a mix u e o MTS, MTD and TST (2.5610–6 mol/L o each one in 1% / in me ha- nol:wa e ). Expe imen al condi- ions as in Fig. 4.