capillary zone electrophoresis in the forensic toxicology ... · for capillary electrophoresis v o...

The worldwide newsletterfor capil lary electrophoresis

Volume 10, Issue 3 • December 2006

GENERAL

After twelve years, the use ofCapillary Zone Electrophoresis(CZE)1-4 is well entrenched in

the day-to-day validated analyticalprocesses in Forensic Toxicology Ser-vices in the Royal Canadian MountedPolice Forensic Laboratories. Wholeblood and urine is reliably screenedin automated, batched casework.Specificity for unknown peaks inthe screening process is en-sured by coupling highlyprecise mobilities with UVspectra. This is achieved throughthe use of databases (libraries) andsearchable lists (called Peak/GroupTables) of drug and drug metabolites.This combination of parameters,when applied to a process of elimina-tion of a large number of compoundsin an analytical database, permits the

analyst to reduce the number ofpossibilities for the unknown to asmall and manageable number, oftenyielding the exact identification. In

many cases, required confirmation byanother hyphenated technique likeGC/MS or LC/MS/MS is almostredundant.

CAPILLARY ZONE

ELECTROPHORESIS IN THE

ANALYTICAL PROCESS

The best way to illustratethe usefulness of CZE in Foren-sic Toxicology is to look at theanalytical process involved(Figure 1), with the concepts,aspects and relevance of CZEin mind at each step. This willbe a process that is very fami-liar to analysts in most analyti-cal disciplines.

SELECTION OF MATERIAL FOR

ANALYSIS

In the evolution of sample prepa-ration in Forensic Toxicology, thegoal has been to be able to analyzewhole blood samples, seized in case-

work, for drugs of forensic signifi-cance. Drug levels in blood givethe closest estimation of effect onthe Central Nervous System (CNS).Design of analytical methods has

always been based on the expectedtherapeutic levels of drugs in blood,which are typically in nanograms permilliliter for the largest group of

significant drugs, the nitrogenousbases. The amount of drug in

other submitted material is almostcertainly higher. Urine levels are anorder of magnitude higher, typicallyin micrograms per milliliter. Tissuesamples are similarly higher thanblood by at least an order ofmagnitude or more. Other samplessubmitted, powders or liquids maycontain drugs at much higher levels.Finding drugs at milligrams per gramof material or much higher is notuncommon.

The goal, then, is to be able to de-tect forensically significant levels ofdrugs and their metabolites in bloodsamples. As a result, the detection ofanalytes in other material is simply amatter of using less material foranalysis or dilution of the extracts.

Capillary Zone Electrophoresis in the Forensic Toxicology ProcessJOHN C. HUDSON, M.SC., FORENSIC TOXICOLOGIST (RETIRED)R.C.M. POLICE, FORENSIC LABORATORY, REGINA, SASKATCHEWAN, CANADA

Figure 1. The toxicology analytical process.

2


EXTRACTION PROCEDURES

After you have selected the typeof exhibit material, it is necessaryto do some kind of preconcentrationusing an extraction procedure.

Many types of extraction proce-dures are available and selection isbased primarily on the matrix invol-ved. Liquid-Liquid Extraction (LLE)is a commonly used procedure world-wide for drugs encountered in Foren-sic Toxicology. Reasons for thecontinued use of LLE as the primarymode of sample preparation aremany, and include low cost, cleanerextracts, and simplicity in laboratorytraining and application. Solid PhaseExtraction (SPE) is used when auto-mated high throughput is desired,but is expensive and extracts fromwhole blood are not usually as cleanas those from LLE. Other material,urine for example, can be extractedusing LLE or proprietary proceduressuch as Toxi-Lab (Varian, Inc.) withgood results. Analysis of solids orliquids is straightforward and usuallyinvolves a simple dilution step beforeeach run.

EXTRACTION PROTOCOL FOR

BASIC DRUGS

The protocol for LLE of basicdrugs is straightforward and consistsof 1–chlorobutane extraction ofblood samples to which an internalstandard has been added.5

TO 1 ML OF WHOLE BLOOD:

1. Add 0.2 mL conc. NH4OH, vortex,and add 5 mL of 1–chlorobutane.

2. Shake for 10 minutes on a flatbedshaker, centrifuge at -10°C for10 minutes at 3000 rpm.

3. Transfer the organic layer to a12x75 mL tube and evaporate to~0.5 mL.

4. Add 10 µL of 1% HCl in MeOH(fresh monthly). Vortex andevaporate to dryness.

5. Add 30 µL water to tube, vortex,and heat briefly to dissolve.

6. Centrifuge at high speed—12,000 rpm for 20–60 minutes.

7. Inject via voltage, 16 seconds at10 kV.

SAMPLE PREPARATION

Before any actual instrument runscan be done, the sample residuemust be prepared for the CZEanalysis and applicable injectionmode. The CZE systems are aqueousin nature and this must be taken intoconsideration in the treatment of thefinal residue.

It is important to add a smallamount of 1% HCl in Methanol to theremaining 0.5 mL of 1–chlorobutanein the evaporations stage of the pro-tocol. The HCl makes the hydro-chloric acid salt of any basic drugspresent, increases the solubility ofthe drug in water, and prevents lossin the final evaporation stage.

The residue is then dissolved byheating in the best-quality wateravailable and centrifuged at highspeed with an angled rotor, toremove particulate matter that mightblock the capillary.

Figure 2. Quality control instrument standard run daily.

SAMPLE INJECTION

Electrokinetic (voltage), ratherthan pressure injection mode, shouldbe used on these samples. Voltage in-jection provides a ‘built-in’ cleanupof dirty samples commonly encoun-tered in Forensic Toxicology case-work. The cations of interest areselectively introduced onto thecolumn, leaving neutral and anionic(acidic) material behind.

This pre-concentration step pro-vides between a 3–10 time sensitivityenhancement over pressure injec-tions and is required for detection ofbasic drugs at nanogram-per-milliliterlevels.

3

Figure 3. Confirmation by the coupling of Mobility, UV Spectrum, and MS/MS Data.

A

B

Figure 4. Cyclodextrin-aided separation of a complex mixture of common drugs.

4

RUN CONDITIONS

Sample runs are done at low pHunder the conditions describedbelow:

• 75 µm ID, 375 µm OD bare fused-silica column

• 50 cm to detector; 60.3 cm total

• Buffer 100 mM Phosphate

• pH 2.38 +/- 0.02

• 12 second, 10 kV injections ofquality control standards, 5 ng/µLof each in 10 mM Buffer

• 16 second, 10 kV injections ofeach case sample in 30 µL ofwater

QUALITY CONTROL RUNS

A typical Quality Control (QC)standard run before each batch ofcases is shown in Figure 2. Thiselectropherogram provides visualdiagnostic information prior to eachbatch. The horizontal line towardsthe top of the trace is the currentflowing in the column during therun and must be flat and stablethroughout each run. All 18 peaksmust be present, above a specifiedabsorbance and must be correctlyidentified.

Extracts of 1 mL blood samples,spiked with 10 nanograms per milli-liter of each component, provideevidence of suitable sensitivity dur-ing each batch run.

AUTOMATION

Sequence (batch) tables allowunattended collection and processingof the QC and Case Samples.

After the runs have been acquir-ed, the software processes each run,first by finding the Internal Standard(IS), then calculating the mobility ofeach peak relative to the referencemobility of this standard. The mobil-ity of each peak, in combination withits UV spectrum, is compared to aPeak/Group Table, which providessuggested peak identification(Figure 3a).

Depending on the quality ofthe results, reruns may be requiredfor samples that are not suitable. Forexample, dilutions of complex andoverloaded samples may benecessary before the results can beinterpreted.

Problem samples, such as verycomplex mixtures, can be rerun ineither or both of two well-defined cyclodextrin systems. Cyclodextrinsare potential chiral or complexing ad-ditives to existing buffer systems.

As an example of the power ofthe use of cyclodextrin additives inroutine casework, Figure 5 showssimilar standards on commonlyencountered drugs run in the twomain, complementary systems

Diphenhydramine, lidocaine, andits main metabolite, MEGX are notseparated well in the 2.38 System, butare resolved by more than 12 min-utes using the 1.2% ß–Cyclodextrinin pH 2.38 phosphate buffer.

INTERPRETATION AND

VALIDATION OF THE SCREEN

A difficult question for any Foren-sic Toxicologist to answer is, “On anygiven day, how good a job are you do-ing in detecting drugs?”

It is simply not practical to suggestthat all analytes of interest can be runas controls to satisfy casework QC re-quirements. Attempts to do this inthe past have typically included run-ning representative compounds withdifferent functional groups or drugsfrom varying pharmacological clas-ses. Answering the question usuallyinvolves reference to old data runswhere similar compounds weredetected.

Up to this point, the difficultywith traditional chromatographicmethods of analysis included adsorp-tion of compounds on active sitesand carry-over problems.6 The useof CZE eliminates these concernsand even allows the analyst to use‘old data’ for quantitative estimates.

The Copley thesis4 is really thefirst attempt to define and addressthe problems encountered with GCand HPLC methods. In this study,fifty-two drugs were validated fromboth a qualitative and quantitativestandpoint by CZE.

It was shown that the validationdata can be used to provide confi-dence in the screen and permit avery close estimate of validation para-meters from the data. This provides


Figure 5. Electropherogram from impaired driving case.

5

good data on which to design a fullquantitation in routine casework andeliminates the test run before the fullcalibration.

CONFIRMATION ISSUES

Unequivocal confirmation is anecessity in Forensic Toxicology. Inthe litigation process, it is a must thatthe compound of interest be thor-oughly confirmed, preferably using a‘gold standard’ confirmation tech-nique. This invariably means using ahyphenated technique such as massspectrometry, coupled to a front-endseparation technique such as GC orHPLC.

Even though the use of mobilityand UV spectra is really an exampleof a hyphenated technique, confirma-tion using mass spectrometry mustbe done.

It is obvious, then, that it wouldbe most efficient to couple CZE/PDA with Mass Spectrometry. Thecombination of mobility, UV Spectra,and Mass Spectrometry will with-stand any legal scrutiny in the future(Figures 3a and 3b).

The Beckman Coulter P/ACEMDQ has been successfully inter-faced to tandem mass spectrometersand ion traps of many major manu-facturers including ThermoFinnigan,ABI/Sciex, Waters, and BrukerDaltonic. The sensitivity of thesemodern mass spectrometers haseliminated the concern about massloading via CE and subsequentsensitivity issues. Excellent resultshave been determined using thishyphenated technique.7

QUANTITATION OF DRUGS IN

SMALL BLOOD SAMPLES

Following drug confirmation, anynecessary quantitations areperformed.

We use a type of UV detector, thephotodiode array detector (PDAD),which is, inherently, a very stabledetector, giving us good linearityover wide concentration ranges.

Samples received in Forensic Tox-icology cases are sometimes verysmall, less than 0.5 mL. In this caseexample, involving fatalities causedby an impaired driver, MDMA (Ecsta-sy), ketamine, and its metabolite,norketamine, were easily detectedand quantitated on 0.25 mL samplereplicates with very good sensitivity.Figure 5 is the electropherogram forone of the small blood samples. Alldrugs were detected at nanogram-per-milliliter levels; MDMA,92 ng/mL; Norketamine, 45 ng/mL;Ketamine, 33 ng/mL.

Regression analysis can beperformed directly on the acquireddata with excellent linearity and easeof processing.

CONCLUSIONRoutine screenings of whole blood

and urine samples, separation of com-plex mixtures using cyclodextrins,and quantitation of the typical smallsamples encountered in forensiccasework, are now routine applica-tions of Capillary Zone Electropho-resis methods and technology.

REFERENCES1. Hudson, J. C., Golin, M., and

Malcolm, M., “Capillary ZoneElectrophoresis In A Compre-hensive Screen For Basic Drugs InWhole Blood” Can. Soc. Forens.Sci. J. Vol. 28, 2 137-152 (1995).

2. Hudson, J. C., Golin, M., andMalcolm, M., and Whiting, C. F.,“Capillary Zone ElectrophoresisComprehensive Screen For DrugsOf Forensic Interest In WholeBlood: An Update” Can. Soc.Forens. Sci. J. Vol. 31, 1 1-29(1998).

3. Golin, M., “Evaluation and Appli-cation of Capillary Zone Electro-phoresis In Forensic Toxicology”A Thesis Submitted to the Collegeof Graduate Studies andResearch in Partial Fulfilment ofthe Requirements for the Degreeof Master of Science in theToxicology Graduate ProgramUniversity of Saskatchewan,Saskatoon. Spring 1999.

4. Copley, H., “Validation of Capil-lary Electrophoresis in ForensicToxicological Analyses: ScreeningWhole Blood for Common BasicDrugs” A Thesis Submitted tothe College of Graduate Studiesand

Research in Partial Fulfilment ofthe Requirements for the Degreeof Master of Science in the Toxi-cology Graduate Program,University of Saskatchewan,Saskatoon. 2004.

5. Sharp, M. E., “Evaluation of aProcedure for Basic and NeutralDrugs: n-Butyl Chloride Extrac-tion and Megabore Capillary GasChromatography.” Can. Soc.Forens. Sci. J., Vol. 19(2), 83-101(1986).

6. Hudson, J. C., Malcolm, M. J., andGolin, M., “Advancements inforensic toxicology: CZE replacesGC/NPD as the screen of choicefor basic drugs” P/ACE SetterNewsletter Vol. 2 Issue 4 1-5(1998).

7. Hudson, J. C., Chen, A. F., andChapman, J. D., “Drug Confirma-tion with CE/MS/MS” P/ACESetter Newsletter Vol. 6 Issue 16-10 (2002).

INTRODUCTION

Assaying for a drug in a biologi-cal sample implies performingtoxicity tests to determine the

presence of a drug administered tothe body. The study of kinetics fora molecule in the body is performedduring investigation or research con-ducted to characterize parameterssuch as absorption, metabolism,distribution, or excretion. In bothcases, a simple method for determi-nation, resulting in high precisionand accuracy as well as high sensitivi-ty, should be employed. This reportoutlines capillary electrophoresisconditions for the analytical charac-terization of some common drugs.

Biological samples used for testinginclude blood (serum and plasma),urine, saliva, amniotic fluid,1 andbiological tissues. In this report, weaddress the use of Capillary Electro-phoresis (CE) as an analytical tech-nique for the analysis of drugspresent in blood samples.

CONCENTRATION MEASURE-MENT OF VARIOUS DRUGS

BY CEWe have developed analytical

methods for drug analysis.

1. ANTIMICROBIAL AGENTS

Many antimicrobial agents havetheir own UV absorption propertiesand can be easily detected, whileaminoglycoside antimicrobials (ami-kacin, kanamycin, etc.), having no

such property, require indirect UVdetection.5 Generally, a peak in anelectropherogram points upward (ina positive direction), while it pointsdownward (in a negative direction)when measured by this indirectmethod, however, the same quantita-tive calculation method can be used.Fluorescence detection6 or electro-chemical detection7 may be used insome cases.

(1) VANCOMYCIN (GLYCOPEPTIDE)8,9

Vancomycin exhibits therapeuticproperties against MRSA and otherbacteria, has a large-ring structure,and can be resolved by CE as it isbasic and shows high polarity. Anexample separation profile andanalytical conditions are shown inFigure 1. For each sample, thecapillary tube is rinsed 3 minuteswith 1.0 M sodium hydroxide,2 minutes with distilled water, andconditioned with the run buffer for5 minutes before being injected witha serum sample directly. Capillaryrinsing is required, following eachsample, to ensure impurities don’tbond along the inner wall of thecapillary tube and to maintain highreproducibility. In addition, periodicrinsing of the capillary tube with1.0 M hydrochloric acid should beperformed for the same reason. Theuse of a fused-silica capillary allowsfor the use of most solvents havingstrong acidity, strong alkalinity, andalso enables freedom from clogging,thus offering high durability. Thedetection limit was 1.0 mg/L (S/N=3),reproducibility (6.3–5.0 mg/L in aver-age with 4.0–5.5% of coefficient ofvariation (CV)) and recovery ratio(91–103%) were good. There was

6


Figure 1. Typical electropherograms of blank human serum (A), serum standard (24.0 mg/L) (B), and serum sample froma patient on vancomycin (50.9 mg/L) (C). Run buffer, borate buffer (0.025 M, pH 10.0) containing 0.1 M SDS; capillary, afused-silica capillary tube (effective length 50 cm; internal diameter, 75 µm); pressure injection time, 4 sec (0.5 psi);applied voltage, 25 kV; capillary temperature, 25°; and UV detection wavelength, 210 nm.

Using Capillary Electrophoresis for Analysis of Therapeutic Drugs in Biological SamplesTOSHIHIRO KITAHASHI, DEPARTMENT OF LABORATORY MEDICINE, KINKI UNIVERSITY SCHOOL OF

MEDICINE, AND ETSUO ARAI, BECKMAN COULTER K. K. TOKYO JAPAN

no interference with the determina-tion from 34 drugs, including otherdrugs and vancomycin metabolites(CDP-1-m and CDP-1-m). As for itscorrelation with other methods, itcorrelated with fluorescence polar-ization immunoassay (FPIA) with thecorrelation coefficients of r=0.982(n=31), y(CE)=0.946x(FPIA)-0.830,and with HPLC with r=0.985 (n=30),y(CE)=0.888x(HPLC)+0.134.

(2) MEROPENEM (CARBAPENEM)10

Meropenem has strong antimicro-bial activity against Gram-negativebacteria including Pseudomonasaeruginosa. This drug can be ana-lyzed by direct serum injection usingMEKC (Figure 2). It is a general prac-tice to increase sample injection timeto enhance assay sensitivity, but aninjection time of 6 seconds or moreby this method caused peak broaden-ing and tailing to occur. Applicationof lower voltage increases sensitivitybut also greatly lengthens migrationtime and broadens peak shape. Themaximum UV absorption for theanalyte is in the range of 200 nmor below, as well as at 297 nm. At200 nm or below, although sensitiv-ity increases in a quantum manner,interference from endogenous sub-stances in serum is sizable, makinganalysis difficult. The detection limitis 2.0 mg/L (S/N=3) and the troughvalue of this drug in the body systemis 4.0 mg/L, which makes this methodclinically applicable. The CV ofreproducibility is 3.43–8.87% (6.3-100 mg/L in average) and the recov-ery ratio 92–111%.

(3) LINEZOLID (OXAZOLIDINONE)11,12

Linezolid exhibits an antimicro-bial property against Vancomycin-resistant intestinal cocci. A run bufferof 50 mM boric acid sodium tetra-borate buffer (pH 8.0) containing50 mM SDS is used and detectionshould be carried out at 250 nm.The migration time is 5.5 minuteswith a limit of detection of 0.5 mg/L(S/N=3). Direct serum injection

analysis often poses a problem ofinterference from high concentra-tions of endogenous substances thataffect the analysis, but with thismethod, there was no influence ofbilirubin 0–41 mg/L, hemoglobin0–914 mg/L, and chyle 0–4940 mg/Lformazin turbidity at their respectiveconcentration range.

(4) FLOMOXEF (OXACEPHEM)13

Flomoxef functions as an antimi-crobial against Gram-negative bacte-ria and also exhibits wide antimicro-bial properties against bacteria ingeneral. Direct serum injection ana-lysis was attempted, but separationfrom endogenous substances inserum such as proteins could not beaccomplished, leading to the needfor pretreatment by deproteinization.A 0.1 mL volume of serum was mixedwith 0.4 mL of acetonitrile containingacetaminophen as an Internal Stan-dard (IS), centrifuged at 2,000 x gfor 2 minutes, after which the super-natant was evaporated to dryness(at room temperature) under reducedvoltage and the resultant substancewas redissolved in 0.05–0.1 mL ofdistilled water, to be analyzed byMEKC. During sample preparation,an IS should be used to make up for

loss of drugs during the pretreatmentsteps. It goes without saying that theIS must be completely resolved fromthe drug itself. Using a 25 mM boraterun buffer (pH 10.0, 50 mM SDS) andthe deproteinization process, detec-tion can be done at 200 nm. Theassay sensitivity was 1.0 mg/L(S/N=3) with no interference withthe determination from othercephem drugs such as cefazolin,cefmetazole, cefoperazone,cefotaxime, and ceftriaxone.

(5) CEFOZOPRAN (CEPHEM)14

Cefozopran has a wide-spectrummode of action with a strong antimi-crobial preference for Gram-positiveand -negative bacteria. Analysis wasperformed using 25 mM sodiumtetraborate 0.1 M sodium hydroxiderun buffer (pH 10.0, 50 mM SDS) bydirect serum injection (detected at244 nm). The limit of detection was0.5 mg/L (S/N=3), while 0.05 mg/Lhas been reported using HPLC. Ifsample pretreatment is performed,an increase in sensitivity can beexpected. Without pretreatment, anincrease in baseline noise level isunavoidable resulting in somedecline in sensitivity.

7

Figure 2. Typical electropherograms of blank human serum (A), serum standard (25.0 mg/L) (B), and serum sample froma patient on meropenem (15.9 mg/L) (C). Run buffer, sodium tetraborate buffer (0.025 M, pH 10.0) containing 0.09 MSDS; capillary, a fused-silica capillary tube (effective length 67 cm; internal diameter, 75 µm); pressure injection time, 5 sec(0.5 psi); applied voltage, 25 kV; capillary temperature, 20°; and UV detection wavelength, 297 nm.

8


(6) CEFEPIME (CEPHEM)15

Cefepime is effective againstGram-positive bacteria. It can beanalyzed with good precision andaccuracy by direct serum injectionusing 100 mM borate run buffer(pH 8.0, 80 mM SDS).

(7) MICAFUNGIN (CANDIN)16

Micafungin is a relatively newantimicrobial agent. Its pretreatmentis performed by adding 0.2 mL ofacetonitrile containing an IS (aceta-minophen) to 0.1 mL of serum fordeproteinization. The supernatantis injected 10 seconds (0.5 psi), at25 kV electrophoresed using 100 mM

borate buffer (pH 9.5) and detectedat 200 nm. The lower the capillarytemperature, the greater the migra-tion time, causing a decline ratherthan an increase in sensitivity. Forthis reason, the analysis should beperformed at 25°C.

2. PILSICAINIDE (AN ANTI-ARRHYTHMIC AGENT)17

Pilsicainide is effective for ventric-ular and supraventricular arrhythmia,but has been attributed to ventricularfibrillation and tachycardia. Its thera-peutically effective concentration inblood is 0.2–0.9 mg/L. As its measure-ment range concentration is low,liquid-liquid extraction is to be em-ployed. The extraction procedureincludes addition of 0.05 mL of anIS (procainamide) solution (10 mg/L)to 0.5 mL of serum, followed by addi-tion of 0.5 mL of 0.1 M sodiumhydroxide for alkalization, extractionusing 5 mL of diethyl ether, evapora-tion of the supernatant, and finallyredissolution of the residue with0.05 mL of distilled water for analysis.Pilsicainide in serum is thus mea-sured as a 10-fold concentratedsample. Its electropherogram isshown in Figure 3. The detectionlimit is 0.04 mg/L (S/N=3), linearity(0–2.0 mg/L) is good and the CV forwithin-run reproducibility is0.798–2.320% (0.25–2.33 mg/L).

3. THEOPHYLINE (AN ASTHMATIC

AGENT)18

Theophyline is analyzed by directserum injection with 20 mM disodi-um hydrogenphosphate-sodiumdihydrogenphosphate run buffer(pH 7.0, 50 mM SDS) and detectedat 270 nm. The migration time is4.7 minutes with a detection sensiti-vity of 1.0 mg/L (S/N=3). This iscorrelated with an enzyme immuno-assay (EMIT) with r=0.973 (n=57),y(CE)=1.01X(EMIT)+0.516.

8

Figure 3. Typical electropherograms of blank human serum (A), serum standard (1.0 mg/L) (B) and serum sample from apatient on pilsicainide (0.57 mg/L) with internal standard (procainamide) (1) and pilsicainide (2) (C). Run buffer, sodium di-hydrogenphosphate buffer (0.1 M, pH 2.29); capillary, a fused-silica capillary tube (effective length 67 cm; internaldiameter, 75 µm); pressure injection time, 16 sec (0.5 psi); applied voltage, 25 kV; capillary temperature, 25°; and UVdetection wavelength, 200 nm.

9

4. IOHEXOL (AN ANGIOGRAPHIC

AGENT)19

Iohexol is an angiographic agentthat has been attributed to serioushealth problems including shock anddecreased kidney function. The CEmethod includes direct serum injec-tion in 50 mM borate run buffer(pH 9.5, 50 mM SDS) and detectioncarried out at 245 nm. The detectionsensitivity is 0.5 mg/L (S/N=3) andprecision and accuracy are good.

SUMMARYCapillary electrophoresis is a

technique that is highly suitable forresearch where only nanoliters ofsample are available. In addition,the low cost of reagents make it anattractive technique, offering excel-lent operability. Additionally, formethod development and routinetesting, use of fused-silica capillarytubes, in comparison to resin-packedHPLC columns, contributes to therobust functionality due to their dura-bility in severe analytical conditions.

REFERENCES1. Stewart, C.J., Iles, R.K., Perrett, D.,

“The analysis of human amnioticfluid using capillary electropho-resis.” Electrophoresis 22. 1136-1142, 2001

2. Reinhoud, N.J., Tjaden, U.R., Irth,H., “Bioanalysis of some anthra-cyclines in human plasma bycapillary electrophoresis withlaser-induced fluorescencedetection.” J.Chromtogr BiomedAppl 574. 327-334, 1992

3. Thormann, W., Lienhard, S.,Wernly, P., “Strategies for themonitoring of drugs in bodyfluids by micellar electrokineticcapillary chromatography.”J. Chromatogr 636. 137-148, 1993

4. Shihabi, Z.K., “Sample matrixeffects in capillary electropho-resis.” J. Chromatogr A 652.471-475, 1993

5. Ackermans, M.T., Everaerts, F.M.,Beckers, J.L. “Determination ofaminoglycoside antibiotics inpharmaceuticals by capillary zoneelectrophoresis with indirect UVdetection coupled with micellarelectrokinetic capillarychromatography.” J. Chromatogr606. 229-235, 1992.

6. Oguri, S., Miki, Y., “Determina-tion of amikacin in human plasmaby high-performance capillaryelectrophoresis with fluorescencedetection.” J. Chromatogr B 686.205-210, 1996.

7. Stead, D.A., “Current methodolo-gies for the analysis of amino-glycosides.” J. Chromatogr B747. 69-93, 2000.

8. Furuta, I., Kitahashi, T., Kuroda,T., et al. “Rapid serum vanco-mycin assay by high-performanceliquid chromatography using asemipermeable surface packingmaterial column.” Clin ChimActa 312. 221-225, 2001.

9. Kitahashi, T., Furuta, I., “Determi-nation of vancomycin in humanserum by micellar electrokineticcapillary chromatography withdirect sample injection.” ClinChim Acta 312. 221-225, 2001.

10.Kitahashi, T., Furuta, I., “Determi-nation of meropenem by capillaryelectrophoresis using directinjection of serum.” J. Chroma-togr Sci (in press).

11.Kitahashi, T., Furuta, I., “Methoddevelopment for determining theantibacterial linezolid in humanserum by micellar electrokineticcapillary chromatography.”J. Pharm Biomed Anal 30.1411-1416, 2002.

12.Kitahashi, T., Furuta, I., “Furthermethod development for mea-surement of linezolid in humanserum by MEKC.” J. PharmBiomed Anal 35. 615-620, 2004.

13.Kitahashi, T., Furuta, I., “Determi-nation of antibacterial flomoxefin serum by capillary electropho-resis.” J. Chromatogr Sci 41.173-176, 2003.

14.Kitahashi, T., Furuta, I., “Develop-ment and validation of a MEKCmethod for the direct determina-tion of cefozopran in humanserum.” J. Pharm Biomed Anal34. 409-414, 2004.

15.Kitahashi, T., Furuta, I., “Methoddevelopment and validation forthe direct determination ofcefepime in human serum bymicellar electrokinetic capillarychromatography.” CurrentPharm Anal 2: 17-21, 2006.

16.Kitahashi, T., Furuta, I., “Analysisof micafungin in serum bycapillary zone electrophoresis.”J. Chromatogr Sci 44. in press,2006.

17.Kitahashi, T., Furuta, I., “Quanti-fication of pilsicainide in serumby capillary electrophoresis.”J Pharm Biomed Anal 29.625-629, 2002.

18.Kitahashi, T., Furuta, I., “Determi-nation of theophylline by micellarelectrokinetic capillary chromato-graphy with direct serum injec-tion.” Jpn J. Clin LaboratoryAutomation 27. 41-46, 2002.

19.Kitahashi, T., Furuta, I., “Methoddevelopment for determining theiohexol in human serum by micel-lar electrokinetic capillary chro-matography.” J. Pharm BiomedAnal 34. 153-158, 2004.

10


LASER-INDUCED FLUORESCENCE

TEST MIX CALIBRATION AND

PERFORMANCE TESTS

Laser-induced fluorescence (LIF)detection differs from absor-bance detection in some impor-

tant ways. An absorbance detectormeasures a small-intensity differencein a high intensity light source. Thisis an effective technique because thepercentage of light does not changeas the intensity of the light sourcechanges. This means peak responsefor a sample will remain constant asthe lamp ages or when a lamp ischanged. With LIF detection, lowlight intensity is measured from adark background. These low intensi-ties are more easily influenced bychanges in the optical path. A testmix calibration can be performed tominimize these influences. For thepurpose of this article, LIF calibrationfor a 488 nm laser with fluoresceinwill be discussed.

RELATIVE FLUORESCENT UNITS

(RFU)LIF detector response is

annotated in units of Relative Fluo-rescent Units (RFU) as opposed to lu-mens or some other unit of lightenergy.

To provide performance specifica-tions for the LIF detector, it was nec-essary to develop a fluorescent testmix solution. The performance testmix was chosen to be 1 x 10-7 MFluorescein and Sodium Salt in water.

THE CALIBRATION CORRECTION

FACTOR (CCF)Several different LIF systems

were used to establish an expectedresponse value for the LIF test mix.For the performance test capillary(60 cm total length, 75 µm diameter,bare fused silica–BFS), the expecteddetector response was determinedto be 35. Most of the systems testedwere within +/-10 % of this value.To correct for this variability, acalibration correction factor (CCF)can be calculated for each system asfollows:

Expected response (Target) = 35

Measured response = M

CCF = 35/M

APPLYING THE CCFIn the 32 Karat™ instrument con-

figuration, LIF calibration is availablefor systems with LIF detection. To ac-cess the LIF calibration wizard, openthe LIF instrument configurationdialog box (Right click on instru-ment, select configuration, selectappropriate instrument type, double-click the LIF detector icon in rightplane. See Figure 1).

MANUAL VERSUS AUTOMATIC

CALIBRATION

Manual calibration must alwaysbe performed as the first step in testmix calibration. If the CCF is known,enter the value and calibration iscomplete. If the CCF is not known,enter 1.0 then perform an automaticcalibration. Automatic calibrationwill display a series of screensexplaining each step of the calibra-tion process. During this process,the instrument runs a method andcalculates a CCF. If this value isbetween 0.1 and 10.0 it should beaccepted to complete calibration.CCF values outside of this rangelikely indicate a problem in theoptical path.

Keeping up the P/ACE: Technical Insight IntoP/ACE MDQ, ProteomeLab™ PA 800, and 32 KaratRICHARD CARSON, BECKMAN COULTER, INC. FULLERTON, CALIFORNIA

Figure 1.

17.5 LIF - Channel 1Test-4

15.0

12.5

10.0

7.5

5.0

2.5

0.0

17.5

15.0

12.5

10.0

7.5

5.0

2.5

0.0

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

RFU R

FU

Minutes

Figure 2. Water and fluorescein in 30 cm total length, 50 µm diameter, BFS capillary.

11

THE CALIBRATION METHOD

Automatic calibration can seem abit mysterious because the methodparameters are not shown and nodata is displayed. The method isquite simple. The capillary is firstrinsed with a non-fluorescent solu-tion. Next, a pressure separation isperformed without voltage. Data iscollected during this step. Finally, thecapillary is rinsed with more of thenon-fluorescent solution to purge thetest mix solution. The result is a step-shaped data file (see Figure 2).

NON-FLUORESCENT SOLUTIONS

Typically, the non-fluorescentsolution used in the calibration is theseparation buffer. Always be carefulthat this solution is safe for use in thetype of capillary being used. Anotherconcern with buffers is possibleinteraction with fluorescein. Whenborate and fluorescein are mixed, thefluorescence increases for a short pe-riod of time. This will cause a bumpin the step data. To eliminate thebump, the fluorescein is mixed 1:1with the buffer for the calibrationmix. This mixture loses fluorescenceafter an hour or more and should bediscarded. Water is almost alwayssafe on capillaries and does not inter-act with fluorescein (do not need tomix 1:1 for test solution). An expect-ed target response value for water/fluorescein calibration has not beenestablished for the 60 cm, 75 µmdiameter, bare fused-silica capillary(Water does not make as good aseparation buffer. See performancetests below).

LIF CALIBRATION FOR DIFFER-ENT DIMENSION CAPILLARIES

For detector performance tests, a60 cm total length, 75 µm diameter,bare fused-silica capillary is used.Many applications require capillarieswith different dimensions. This canaffect the test mix calibration in twoways. If the capillary length is differ-ent, the step change will occur at adifferent time. Automatic calibrationcalculates this time based on thevalues entered for the capillary in thecalibration wizard. If the diameter ofthe capillary is changed, the targetvalue will be different. For a barefused-silica capillary, this new targetvalue can be calculated based on the75 µm capillary target:

75 µM CAPILLARY CROSS-SECTIONAL

AREA:= p • r^2 = 3.14 • (75/2) ^ 2= 4418

50 µM CAPILLARY CROSS SECTIONAL

AREA:= 3.14 • (50/2) ^2= 1963

AREA RATIO:= 1963/4418= .44

50 µM TARGET:= .44 • 75 µm target= .44 • 35 RFU= 15.4

The calibration wizardrequires an integer for atarget value, so the new50 µm bare fused fluores-cein target is rounded to15. This rounding is not aproblem because RFU arearbitrary units. As long asthe same number is usedas the target, a given con-centration should havethe same response on anydetector, capillary, or laser.Because the target isrounded down, it is likelythe raw detector response

will be slightly higher than the targetvalue. This will cause the CCF to beslightly smaller than one. This willnot have any effect on the sensitivitybecause the signal and the noise arecorrected by the same CCF.

LIF CALIBRATION FOR DIFFER-ENT FLUORESCENT SOLUTIONS

A given application may requireuse of a special fluorescent marker.Depending on the marker, the emis-sion wavelength may be significantlydifferent than fluorescein. In thiscase, it is likely that different filterswill be required. With different fil-ters, the standard fluorescein targetis no longer valid. Even if fluoresceincan be detected at the new wave-length, a new target value should beestablished. It may also be necessaryto use the application fluorescentsolution instead of fluorescein. Estab-lishing a new target value is simple.Use manual calibration to set theCCF to 1.0. Run the pressure sepa-ration method described above for

3 4 5 6

0 14 7 6

20427

0.3

0.2

0.1

0.0

-0.1

LIF - Channel 1Fluorescein

S/N (ASTM)

RFU

0.3

0.2

0.1

0.0

-0.1

Minutes

USE FLASH-LIGHT TO

VERIFY FIBER OPTIC OUTPUT

VERIFY LASERFIBER OPTICBUSHING IN

INTERCONNECTMODULE

INSPECTCARTRIDGE

PLUG ASSEMBLY

VERIFYCORRECTFILTERS

INSTALLED

VERIFY TESTMIX IS FRESH

INSTALL LASERIN OTHERCHANNEL

VERIFY FLATBASELINE WITHSMALL STEP AT

5 MINUTES

RUN SEPARA-TION WITH

LASER OFF FOR 5 MINUTES

FOLLOWED BY LASER ON FOR

5 MINUTES

RECALIBRATE

END

START

NO

NO

INSTALL OPCAL WITH PLUG

The remaining steps can be used to check noise, drift, and baseline stability

VERIFY INTEGRATIONPARAMETERS

ARE OPTIMIZED

CHECK OPTICS

S/N > 10,000?

RUN SYSTEMPERFORMANCE

TEST

YES

YES

0.1<CCF<1.0

PERFORM TEST MIX

CALIBRATION

Figure 4.Figure 3.


six times. Average the response ofthe step change in the six runs. Thisvalue will be the new target.

If an application requires a coatedcapillary, borate solution should notbe used. Substituting water for theborate should be safe for most capil-lary coatings. When water is used inplace of buffer, do not mix the fluo-rescein 1:1 with buffer or water.Simply inject the fluorescein withoutdilution.

STANDARD SYSTEM PERFORMANCE

TEST

Create a method for a 60 cm totallength, 75 µm diameter, bare fused-silica capillary.

Where the rinse and separationvials contain borate and the injectvial contains a 1:1 mixture of fluo-rescein and buffer, a peak will bedetected at about 5 minutes.Optimize the integration so only themain peak and small contaminantpeaks are detected. Annotate the

peak with S/N (ASTM). On acalibrated instrument, the signal-to-noise ratio should be > 10,000(Figure 3).

TROUBLESHOOTING

Many variables can affect the qua-lity of an LIF method. To simplifytroubleshooting, it is necessary toisolate and verify the performanceof each of these variables. Figure 4shows a list of steps for troubleshoot-ing an LIF system.

Developing innovative solutions in Systems Biology.

Beckman Coulter, Inc. • 4300 N. Harbor Boulevard, Box 3100 • Fullerton, California 92834-3100Sales & Service: 1-800-742-2345 • Telex: 678413 • Fax: 1-800-643-4366 • www.beckmancoulter.com

Worldwide Biomedical Research Division Offices:Australia (61) 2 9844-6000 Canada (905) 819-1234 China (86) 10 6515 6028 Eastern Europe, Middle East, North Africa (41) 22 994 07 07 France 01 49 90 90 00 Germany 49 21 513335 Hong Kong (852) 2814 7431 / 2814 0481 Italy 02-953921 Japan 03-5404-8359 Mexico (55) 56057770 Netherlands 0297-230630 Singapore (65) 6339 3633 South Africa/Sub-Saharan Africa (27) 11-805-2014/5 Spain 91 3836080Sweden 08-564 85 900 Switzerland 0800 850 810 Taiwan (886) 2 2378 3456 Turkey 90 216 309 1900 U.K. 01494 441181 U.S.A. 1-800-742-2345

NL-10321A-B2006-7470-CB-8 ©2006 Beckman Coulter, Inc. Printed in U.S.A. on recycled paper.

* Sciex is a registered trademark of Sciex Inc., Toxi-Lab is a registered trademark of Varian Inc. All other trademarks are the property of their respective owners.

capillary zone electrophoresis in the forensic toxicology ... · for capillary electrophoresis v o...

Documents