ORIGINAL PAPER

Detection of electrophoretically separated cytochromes P450by element-labelled monoclonal antibodies via laser ablationinductively coupled plasma mass spectrometry

Peter H. Roos & Arunachalam Venkatachalam &

Andreas Manz & Larissa Waentig &

Christina U. Koehler & Norbert Jakubowski

Received: 31 March 2008 /Revised: 6 June 2008 /Accepted: 10 June 2008 / Published online: 15 July 2008# Springer-Verlag 2008

Abstract Numerous structurally and enzymatically similarcytochromes P450 (CYPs) are involved in the metabolism ofxenobiotics and are present in different amounts and withdifferent enzyme profiles in human tissues and cells. Analysisof their adaptively regulated and individually variable patternsis a peculiar analytical challenge. We developed a laserablation inductively coupled plasma mass spectrometry (LA-ICP-MS) based method for concomitant detection andsemiquantitative determination of electrophoretically separat-ed and blotted CYPs. The first results are given here for thetwo enzymes CYP1A1 and CYP2E1. Specific monoclonalantibodies directed against the enzymes were differentiallylabelled with europium via a covalently linked chelator andwith iodine, respectively. Analysis of the modified antibodiesshows that both europium and iodine are coupled to the heavyand the light chains of the antibodies. Also, the antibodiesmaintained their antigen-binding properties after labelling asdemonstrated by LA-ICP-MS-analysed immunoblots. Themethod allowed us to detect specifically and concomitantlyboth CYP enzymes in complex biological samples, i.e.microsomes of rat liver and minipig duodenum, which arecharacterized by different levels and proportions of the twoCYP enzymes. A strong CYP1A1 signal is found in livermicrosomes of 3-methylcholanthrene-treated rats, while it is

(nearly) absent in liver microsomes of rats treated withisonocotinic acid hydrazide (isoniazid). The constitutivelyexpressed CYP2E1 is found in microsomes of both treatmentgroups. Duodenal microsomes of minipigs orally exposed topolycyclic aromatic hydrocarbons show a clear CYP1A1signal. Low levels of CYP2E1 can also be detected in thesemicrosomes. The LA-ICP-MS method allows concomitantdetermination of CYPs, thereby exhibiting sensitivity similarto that of conventional chemoluminescence detection viaperoxidase-labelled secondary antibodies. The latter methodallows readout of a single CYP protein in a 1D separation.Although the results presented here are only for labelling byuse of the elements iodine and europium, the same strategycan be applied also for other lanthanide elements incombination with chelating compounds, so LA-ICP-MS ofwestern blots offers a new capability to be applied for highlymultiplexed CYP determinations via labelled antibodies.

Keywords Laser ablation inductively coupled plasma massspectrometry . Cytochrome P450 . 1, 4, 7, 10-Tetraazacyclododecane- 1, 4, 7, 10-tetraacetic acid .

Iodination . Europium . Electrophoresis

Introduction

Cytochromes P450 (CYPs) belong to the group of haemthiolate proteins and are enzymes which are involved in themetabolism of a variety of xenobiotics and also ofendogeneous compounds such as steroids and fatty acidderivatives. The attribute ‘450’ in the name of the enzymefamily is delineated from the position of the Soretmaximum in the absorption spectrum of the reduced Fe2+-CYP-CO complex at 450 nm [1]. The human genome

Anal Bioanal Chem (2008) 392:1135–1147DOI 10.1007/s00216-008-2242-2

A. Venkatachalam :A. Manz : L. Waentig :N. Jakubowski (*)ISAS—Institute for Analytical Sciences,P.O. Box 10 13 52, 44013 Dortmund, Germanye-mail: jakubowski@isas.de

P. H. Roos : C. U. KoehlerIfADo—Leibniz Research Centre for WorkingEnvironment and Human Factors,Ardeystr. 67,44139 Dortmund, Germany

encodes 57 CYP proteins, but the number of known CYPenzymes exceeds 1,000 when also invertebrate, plant andbacterial enzymes are considered [2]. All CYP enzymesexhibit similarity in their structure and general mechanismof action; however, there are significant differences in thedetailed function of individual enzymes as well as in thestructures and properties of their active sites.

In mammalian species, including man, CYPs are themost important class of enzymes involved in the metabo-lism and detoxification of xenobiotics. Their primary task isto introduce new functional groups into a molecule, therebyfacilitating the subsequent work of phase II enzymes whichattach hydrophilic molecules such as glutathione, glucu-ronic acid or sulphate to the CYP-derivatized xenobiotics.This concerted action of enzymes aims to increase the watersolubility of compounds to achieve their efficient excretion.

Roughly, half of the known human CYP enzymesrecognize xenobiotic substrates. These xenobiotics metabo-lizing CYPs are predominantly localized in the membrane ofthe endoplasmic reticulum. There are large similaritiesbetween the different CYP enzymes in their structure and insome cases also with respect to substrate specificity, sodistinction by electrophoretic behaviour, enzymatic reactivityand spectroscopic properties can be a problem, but highlyspecific structural features of individual CYP enzymes can berecognized by antibodies, in particular by monoclonal anti-bodies. However, cross-reactivities of antibodies recognizingmore than one CYP enzyme occur frequently [3].

It is a great challenge to identify and quantify thenumerous different CYP proteins of microsomal fractions.CYP profiles differ between species, between individuals,between tissues and as a function of age. Furthermore, theprofiles are adaptively modulated by xenobiotics, whichmay induce or suppress a set of CYP enzymes dependingon the compound.

The latter is also true for the CYP enzymes dealt with in thispaper. CYP1A1 (enzyme no. 1 in subfamily A of family 1) isnot present or is found at low levels only in several tissuesunder normal conditions. The enzyme can be induced,however, via a receptor-dependent process by certain xeno-biotics, such as polycyclic aromatic hydrocarbons (PAHs),dioxin and coplanar polychlorinated biphenyls [4, 5]. The PAH3-methylcholanthrene is a classic inducer of CYP1A1 [6–8]and was also used in the experiments described here. It hasbeen shown that CYP1A1 expression increases upon oralPAH exposure in many tissues, including small intestine,liver, kidney, lung and spleen of rats and minipigs [9, 10]enabling the respective tissues to metabolize just the inducingPAH compounds. This CYP1A1-dependent PAH metabolismleads to chemically reactive intermediates [11] and isconsidered one of the initial steps in chemically inducedcarcinogenesis by PAH [12, 13]. Induction of CYP1A1 and ofsome other CYP enzymes (CYP1A2, CYP1B1, CYP2S1) is

controlled by the aryl hydrocarbon receptor, a ligand-activatedtranscription factor [14, 15] which is initially localized in thecytoplasm, translocates into the nucleus after ligand bindingand activates transcription after dimerization with arylhydrocarbon receptor nuclear translocator protein [16].

In contrast to CYP1A1, the CYP isoenzyme CYP2E1metabolizes smaller and also hydrophilic molecules. Amongthese are industrial bulk materials such as benzene, toluene,vinyl chloride and trichloroethylene [17]. Other CYP2E1substrates are ethanol and nitrosamines. As in the case ofCYP1A1-dependent PAH metabolism, the latter are con-verted to chemically reactive compounds by CYP2E1 [18].In the liver, CYP2E1 is constitutively expressed, but its levelcan be further increased by various chemicals such asethanol, pyrazol and the tuberculostatic isoniazid [19, 20].Regulation of CYP2E1 expression and enzyme activity israther complex and cannot be discussed here in detail. It isonly mentioned that DNA methylation in the CYP2E1promoter region, ligand-dependent messenger RNA (mRNA)and protein stabilization as well as protein kinase Adependent phosphorylation of the CYP2E1 protein play arole besides transcriptional regulation via cis-acting elements[21–24].

Our approach is to analyse CYP profiles and to quantifyindividual CYPs via element-labelled specific monoclonalantibodies by means of laser ablation (LA) inductivelycoupled plasma (ICP) mass spectrometry (MS). The chiefattraction of this method is that the antibodies can bedifferentially labelled with elements so that target-protein-specific signal discrimination is possible, a prerequisite formultiplexing. Hu et al. [25] successfully applied thisprinciple to detect several proteins by use of samplemicroarrays. For the specific detection of CYP1A1 andCYP2E1, we developed monoclonal antibodies directedagainst the two proteins and developed strategies forlabelling these antibodies using commercially availablechemicals. Protein labelling with lanthanides bound to theprotein via covalently linked chelators is a commonly usedmethod in proteomics [26–30] and has also been applied incombination with ICP-MS [31–33] as well as LA-ICP-MS[25]. In addition to the use of this method for labelling ofantibodies, we provide the first examples for the applicationof iodinated antibodies for the LA-ICP-MS approach.

Experimental

Chemicals

The chemicals and buffers used for the horizontal polyacryl-amide gel electrophoresis (PAGE) labelling reaction weredescribed in previous papers [42, 43]: 2-(4-isothiocyanato-benzyl)-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic

1136 P.H. Roos et al.

acid (p-SCN-Bn-DOTA; Macrocyclics, Dallas, USA);sodium hydrogen carbonate, sodium carbonate, ammoniumacetate, ammonia, mercaptoethanol and bromophenol blue(Merck, Darmstadt, Germany); tris(hydroxymethyl)amino-methane (Tris) hydrochloride, sodium dodecyl sulphate(SDS) and glycine (Applichem, Darmstadt, Germany); 1,4-dithio-DL-threitol (DTT) (Roth, Karlsruhe, Germany);europium chloride (EuCl3∙6H2O) and isonicotinic acidhydrazide (isoniazid) (Sigma-Aldrich Chemie, Munich,Germany); 3-methylcholanthrene and sodium iodide(Fluka, Buchs, Switzerland); acrylamide and bisacrylamide(Bio-Rad, Germany); ammonium peroxodisulphate andglycerol (Roth, Germany); N,N,N′,N′-tetramethylethylene-diamine (Riedel-de Haën, Germany); SDS (Biomol,Germany); and milk powder (GE, Munich, Germany) usedfor blocking the blot membrane. All solutions wereprepared using ultrapure water (Milli-Q water purificationsystem, Millipore, Bedford, MA, USA).

Monoclonal antibodies

For development of monoclonal antibodies directed againstCYP1A1 and CYP2E1, the two proteins were isolatedaccording to previously described chromatographic proce-dures [36, 37] from liver microsomes of rats treated with 3-methylcholanthrene [38] or isoniazid [37], respectively. Micewere immunized with the protein preparations according tostandard procedures until titres of specific antibodies in theserum were sufficiently high. Specific antibody titres weredetermined by an enzyme-linked immunoadsorbant assay(ELISA) applying serial serum dilutions to antigen-coatedELISA sheets. Subsequently, spleen and lymph node cells ofthe mice were isolated and fused with cells of the myelomacell line X63-Ag8.653 [39]. The resultant hybridoma cellswere diluted and disseminated in ten 96-well plates forcultivation resulting in 960 minicultures. Selection ofpositive clones secreting the desired specific antibodies wasdone by ELISA as described by Roos et al. [34] but usingTriton X-100 solubilized rat liver microsomes as antigens.The application of microsomes specifically enriched in theCYP of interest as the antigen was necessary owing to thelarge number of cell clones to be screened, i.e. more than1,000 in each case, for which sufficient amounts of purifiedCYP proteins were not available. The selected monoclonalantibodies used are specific for CYP1A1 and CYP2E1. Inparticular, the CYP1A1 antibody does not cross-react withthe similar CYP1A2 protein.

Production of larger amounts of monoclonal antibodies wasperformed in small miniPERM bioreactors (Greiner Bio-One,Frickenhausen, Germany) as described by Falkenberg et al.[40]. The antibody-containing cell culture supernatant washarvested and kept frozen at −18 °C for further use orprocessing.

Purification of antibodies

Antibodies were isolated and purified from the hybridoma cellculture supernatant by affinity chromatography as follows. Thesupernatant was centrifuged for 10min at 12,000 g and 4 °C toseparate particulate matter. For buffer exchange 2.5-mLaliquots of the clear supernatant were applied on a PD-10column (GE Healthcare, Munich, Germany) which wasequilibrated with 20 mmol L−1 sodium phosphate buffer,pH 7.0, followed by elution with 3.5 mL of the same buffer.The antibody-containing eluate was then applied to a ProteinG-HiTrap column (GE Healthcare, Munich, Germany), whichspecifically binds the antibodies via their Fc. For equilibrationand washing after application of the antibody to be purified,20 mmol L−1 sodium phosphate buffer, pH 7.0 was used.Elution of the antibodies was achieved with 0.1 mol L−1

glycine-HCl, pH 2.7. It is important to collect the antibodybeing eluted directly in 1 mol L−1 Tris-HCl buffer, pH 9.0 forstabilization.

Protein quantification

The concentrations of proteins and antibodies were deter-mined according to the method of Lowry et al. [35] withbovine serum albumin as a standard. For quantification ofantibodies, a Nanodrop instrument (Nanodrop Technolo-gies, Wilmington, USA) was used with immunoglobulin asa protein standard. A 2-μL aliquot of the antibody samplewas used to determine the concentration.

Preparation of microsomes

Rat liver microsomes enriched in CYP1A1 and CYP2E1 wereobtained from animals treated with 3-methylcholanthrene andisoniazid, respectively, as typical inducers of the two enzymes.3-Methylcholanthrene, solubilized in corn oil, was applied bydaily intraperitoneal injections of 20 mg 3-methylcholanthreneper kilogram of body weight over three consecutive days.Induction by isoniazid was achieved via the drinking waterwhich was supplemented by 0.1% isoniazid and which wasaccessible ad libitum for 10 days. On the day following theexposure, the rats were killed, and liver microsomes wereprepared as described by Guengerich [41] and stored at−80 °C. Microsomes from duodenum of minipigs wereprepared as described by Roos et al. [10]. The experimentswere performed in compliance with the German AnimalProtection Law (permissions 23.8720 no. 24.7 and 23.8720no. 20.35, Bezirksregierung Arnsberg).

Antibody labelling with europium and p-SCN-Bn-DOTA

For labelling of the monoclonal CYP1A1 antibody witheuropium we chose p-SCN-Bn-DOTA as a chemically

Detection of electrophoretically separated cytochromes P450 1137

activated chelating agent which reacts with the ε-aminogroups of lysine residues under the pH conditions used.The labelling conditions follow a modified protocol(N. Jakubowski, L. Waentig, H. Hayen, A. Venkatachalam,A. von Bohlen, P.H. Roos and A. Manz, unpublished results)except that the molar ratio of antibody to DOTA to europiumwas maintained at 1:20:200. Briefly, p-SCN-Bn-DOTA wascoupled to the antibody in the first step. The reaction wasperformed in sodium carbonate-bicarbonate buffer, pH 9.0 at20 °C for 80 min. Usually, about 0.4 mg antibody was used forlabelling, corresponding to about 2.7 nmol antibody andDOTA was added in 20-fold excess (54 nmol). After thereaction, unreacted DOTA was removed by size-exclusionchromatography using PD10 columns (Amersham, Freiburg,Germany). In a second step, the antibody-DOTA complex wasloaded with europium. The reaction was performed inammonium acetate buffer, pH 7.4, at 37 °C for 30 min withan excess of europium chloride (540 nmol). Finally, non-bound europium was separated from the antibody-DOTA-europium complex by size-exclusion chromatography onPD10 columns.

Antibody labelling by iodination

The monoclonal antibody directed against CYP2E1 waslabelled with iodine by using Iodobeads (Perbio Science,Niederkassel, Germany) as a commercially available iodin-ation reagent. Briefly, one bead was washed for 30 s in 500 μLof 0.1 mol L−1 Tris-HCl, pH 7.5 in Eppendorf tubes and driedon filter paper. Subsequently, the bead was loaded into a newEppendorf tube together with 250 μL of 0.1 mol L−1 Tris-HCl (pH 7.5), 10 mmol L−1 sodium iodide (N. Jakubowski, J.Messerschmidt, M. Garijo Anorbe, L. Waentig, H. Hayen andP.H. Roos, unpublished results). Incubation took place for5 min usually at room temperature. The production of freeiodine can easily be observed by the change of the colourlesssolution to brown in less than 15 s. In the next step, a fixedamount of 500 μg antibody was dissolved in 250 μL0.1 mol L−1 Tris-HCl, pH 7.5 and added to the reactionvessel. The total reaction volume of 0.5 mL was always keptconstant. The reaction was stopped after 4 min by removal ofthe Iodobead. Each bead was only used once.

Sodium dodecyl sulphate polyacrylamide gelelectrophoresis

Proteins were separated by horizontal SDS-PAGE using aMultiphor II flatbed electrophoresis system (AmershamBiosciences, Freiburg, Germany) connected to the powersupply (2197, LKB, Germany). The handling of thepolyacrylamide gels was described in detail in a previouspaper [42]. Precast CleanGels (T=10%) were purchasedfrom Electrophorese Technik (ETC, Kirchentellinsfurt,

Germany). The electrophoresis running conditions, thewhole protocol, together with the preparation of the rehydra-tion, electrode buffer and sample buffer solutions areexplained in the protocols of Amersham Biosciences forSDS-PAGE (71–7141–00, edition AB). The buffers used forthe SDS-PAGE separation were as follows: rehydration buffer0.3 mol L−1 Tris/acetate, 0.1% SDS, 0.01% sodium azide(pH 8.0); anode buffer 0.3 mol L−1 Tris/acetate, 0.1% SDS,0.01% sodium azide (pH 8.4); cathode buffer 0.8 mol L−1 N-tris(hydroxymethyl)methylglycine, 0.08 mol L−1 Tris, 0.1%SDS, 0.01% sodium azide (pH 7.15); sample buffer 1 mL0.5 M Tris pH 7.5, 0.1 g SDS, 15.5 mg DTT, made upto10 mL with H2O (this buffer was used to dilute themicrosomal protein samples). The gels were reswelled inthe rehydration buffer for 1.5 h. One of the electrode wickswas soaked in anode buffer and the other in cathode buffer.Gel was mounted on the Multiphor II electrophoresis unit,which was cooled by a thermostatic circulator preset at10 °C. The soaked electrode wicks were placed onto theirrespective edges. Aliquots of 0.75–15 μL of each sample andmolecular mass marker were loaded onto the sample wells.The samples were diluted with SDS sample buffer contain-ing the reducing agent DTT and heated to 95 °C for 5 minfor complete denaturation of proteins just before beingapplied to the loading wells. For molecular mass calibration,we used a dual-colour marker covering a molecular massrange from 10 to 250 kDa (161–0374, Bio-Rad, Germany).The marker proteins were iodinated by the same method asdescribed for antibody labelling (see “Antibody labelling byiodination”) to render them detectable by LA-ICP-MS. Theelectrode holder was placed on the electrophoresis unit withthe electrodes resting on the outer edge of the electrodewicks. A lower starting voltage of 65–100 V with a currentof 20 mA was applied initially for 15 min. After that acurrent of 35 mA was maintained with a voltage of 250–300 V for 60–90 min.

Semidry immunoblot

The unstained proteins separated by SDS-PAGE wereimmediately transferred to a blot membrane. Semidryblotting was only performed for unstained gels as previouslydiscussed in a recent paper [43] using the Nova blotting unitfrom Amersham Pharmacia. Proteins were transferred ontoProtan nitrocellulose membranes (Schleicher and Schuell,Dassel, Germany) with a pore size of 0.45 μm. The blottingwas performed using a current of 0.8 mA cm−2 blot area for1 h. The 2197 power supply (LKB, Germany) was used forblotting. For the immunological staining process, eachmembrane was stored overnight in 40 mL of milk powdersolution [5% milk powder in phosphate-buffered saline(PBS)-Tween solution]. The next day the blot was washedwith PBS-Tween solution before being exposed to a

1138 P.H. Roos et al.

solution that contained labelled antibodies that recognize theprotein of interest. Each washing step consisted of three 5-minwashes followed by a 15-min wash with PBS-Tween solution.Labelled antibody solution (with appropriate dilution asprescribed in the product specification list) was preparedusing PBS with a final concentration of 1 μg mL−1. Themembrane was incubated with this solution for 2 h at roomtemperature with shaking. After incubation, the membranewas washed under the same conditions as before with PBS-Tween solution. It was air-dried and then used for LA.

Luminescence-based antibody detection

After blotting, the nitrocellulose sheets were incubatedovernight in PBS (9.5 mM sodium phosphate buffer, pH 6.9,137 mM NaCl, 2.7 mM KCl), 0.1% Tween 20, 10% milkpowder followed by three washing steps with PBS-0.1%Tween 20 for 10 min each. The blot membrane was thenexposed for 1 h to the first antibody (anti-CYP1A1 or anti-CYP2E1) diluted in PBS at a concentration of 1 μg mL−1.After five washing steps with PBS-0.1% Tween 20,incubation with the peroxidase-coupled secondary goatantimouse immunoglobulin G (Sigma, Deisenhofen, Ger-many) was performed for 90 min. Non-bound antibodieswere removed by five washing steps with PBS-0.1% Tween20 (10 min each). Antibody-labelled bands were visualizedby Western Lightning Chemiluminescence Reagent Plus(PerkinElmer, Waltham, USA) and images were acquiredusing the ChemiLux CSX-1400M gel imager (INTAS,Göttingen).

LA-ICP-MS-based antibody detection

For sample introduction into the ICP-MS system, LA wasused. The development of the LA cell used in thisinvestigation and its optimization were described in previouspapers [43, 44].The LA system consisted of a Nd:YAG laser(Minilight II, Continuum, Santa Clara, USA) operated at thefourth harmonic (266 nm) of the fundamental wavelength.

The laser was fired with a constant repetition rate of 15 Hz.A translation velocity of the ablation cell relative to the fixedlaser beam of 1 mm s−1 was chosen to get overlapping craterswith a diameter of 500 μm in the form of a straight line andwas used throughout this investigation.

The nitrocellulose membranes were mounted on a cylin-drical holder and could be rotated by use of a stepper motor sothat any position on the membrane could be reached bytranslation of the cell and rotation of the membrane holder.

The LA cell was coupled to the ICP-MS torch viaflexible polyethylene tubing. Helium was used as atransport gas and argon as a make-up gas. Both gases wereoptimized so as to give maximal signal intensity; for moredetails see [44].

For LA-ICP-MS measurements, a sector-field instrument(Element 2, Thermo Fisher Scientific, Bremen, Germany)was used and the following isotopes were measured in thelow-resolution mode of the mass spectrometer: 153Eu+, 127I+.The instrumental parameters for operation of the ICP-MSsystem are given in Table 1.

The antibody-treated nitrocellullose blot membraneswere rastered for one isotope by line scans with a distanceof 1 mm in-between. Subsequently, another scan for thesecond isotope was run rastering the left 1-mm ridges justin the middle. In some experiments isotopes were measuredsimultaneously in a single run using the updated highelectric scan speed system. Positioning of the laser cell andholder and triggering of the laser was done automatically bya microprocessor control unit. The whole area of themembrane was ablated continuously. The measured inten-sity time profile data of the ICP-MS system was exported tothe program Origin (Additive, Friedrichsdorf, Germany),where intensity-time profiles as well as colour-codedsurface plots can be calculated by transforming time intoa millimetre scale using the translation velocity selected.

Results and discussion

Analyses of labelled antibodies

For evaluation of the labelling success and of maintenanceof antibody integrity, we analysed two element-labelledCYP antibodies, i.e. anti-CYP1A1 and anti-CYP2E1, bySDS-PAGE, blotting and subsequent LA-ICP-MS. Theeuropium-labelled CYP1A1 antibody was analysed bySDS-PAGE under reducing and non-reducing conditionsby inclusion or omission of DTT, respectively. Only oneeuropium-positive band in the correct molecular mass range

Table 1 Instrumental parameters

Laser ablation system

Laser energy 3 mJMake-up gas flow rate (Ar) 1 L min−1

Carrier gas flow rate (He) 1.05 L min−1

Repetition rate 15 HzTranslation velocity 1 mm s−1

Distance between line scans 1 mmInductively coupled plasmamass spectrometry systemIncident power 1,350 WCooling gas flow rate 16 L min−1

Auxiliary gas flow rate 1.3 L min−1

Ni skimmer cone diameter 1 mmNi sampler cone diameter 0.7 mmIsotopes measured 153Eu, 127I

Detection of electrophoretically separated cytochromes P450 1139

of about 150 kDAwas seen under non-reducing conditions,corresponding to the non-dissociated antibody (Fig. 1a).The result indicates successful labelling, maintenance ofantibody integrity upon labelling and purity of the antibodypreparation. Application of different amounts of labelledantibody shows that at least 0.25 μg antibody cor-responding to about 1.67 pmol antibody can be detectedby LA-ICP-MS. Under reducing conditions, SDS-PAGEreveals separation of the dissociated antibody subunits(Fig. 1b). Both heavy and light chains with approximatemolecular masses of 50 and 25 kDa, respectively, werelabelled with DOTA-europium. Labelling degrees related tosubunit size show that the light chain is more efficientlylabelled than the heavy chain. While the molecular massesof heavy and light chains differ by a factor of about 2, the153Eu+ intensities differ only by 1.4. There are severalpossible explanations for the differential labelling efficien-cies, such as different surface to volume ratios and differentnumbers and/or accessibilities of coupling effective lysineresidues. Even under the protein-denaturing conditions ofSDS-PAGE, both in the reducing and in the non-reducingvariant, the europium label remains stuck to the protein.Partial complex dissociation and loss of europium, howev-er, cannot be ruled out. With respect to the immunoblotapplications described later, it is noted that the lanthanide-labelled antibody is faced with milder conditions than forthe electrophoretic separation in the presence of SDS. Thisis an important aspect considering maintenance of theantibody-DOTA-europium complex. The results of theDOTA-europium labelling and the antibody analysis lookvery promising. However, this is not yet proof that theantibody retained its binding activity for CYP1A1. Thiswill be demonstrated later.

The iodine-labelled CYP2E1 antibody was analysed bySDS-PAGE under reducing conditions (Fig. 2). As for theeuropium-labelled antibody, both heavy and light chainswere labelled by the iodination procedure, which involvedtyrosine and possibly histidine residues of the antibodyprotein. The additional bands seen in Fig. 2 can beinterpreted as uncleaved dimers consisting of one heavyand one light chain. For the iodine-labelled CYP2E1antibody, a linear relation between the amount of antibodyapplied and 127I signal intensity is found (r2=0.993, linearregression) (Fig. 3). We plotted the integrated signal of eachlane including all subunits against the amount of antibodyapplied in picomoles in the range from 0.25 μg (1.67 pmol)to 2.0 μg (13.3 pmol). In contrast to chelated europium,iodine is covalently bound to the protein, so no stabilityproblems with the protein label can be expected.

Detection of blotted CYP1A1 with a lanthanide-labelledantibody

To assess the binding activity of the europium-labelled anti-CYP1A1 antibody and its suitability for CYP1A1 detectionby LA-ICP-MS, we applied it in immunoblotting experi-ments. As CYP1A1-containing biological material, we usedliver microsomes of rats which had been treated with theCYP1A1 inducer 3-methylcholanthrene. Their spectro-photometrically determined microsomal CYP contentamounted to 2.08 nmol CYP per milligram of protein.Thus, CYPs make up about 10% of the total microsomalprotein, calculated from an average CYP molecular mass of50 kDa. For separation by SDS-PAGE, we appliedmicrosomal protein in amounts between 0.1 and 5.0 μg.After blotting, detection was done with a europium-labelled

Fig. 1 Analysis of europium-labelled CYP1A1 antibody by laserablation inductively coupled plasma mass spectrometry (LA-ICP-MS)after sodium dodecyl sulphate polyacrylamide gel electrophoresis(SDS-PAGE) and transfer to a nitrocellulose membrane. SDS-PAGE

was performed under non-reducing (a) and reducing conditions (b). Inb, the iodinated molecular mass range marker is shown on the left andthe labelled antibody is shown on the right. The amount of antibodyapplied is indicated. Signal intensity scales are given for each scan

1140 P.H. Roos et al.

monoclonal anti-CYP1A1 antibody by means of LA-ICP-MS (Fig. 4). The 153Eu+ band pattern shows that theantibody specifically recognizes CYP1A1 in the relevantmolecular mass range. The signal intensity corresponds tothe amount of protein applied. The detection limit forCYP1A1 is close to 0.1 μg applied microsomal protein,which corresponds to about 140 fmol CYP1A1. This valueis based on spectroscopic determination of total microsomalCYP and on the fact that CYP1A1 can make up about 70%of the total CYPs in liver microsomes of 3-methylcholan-threne-treated rats [45].

Parallel to detection by LA-ICP-MS, we performedconventional luminescence detection by means of aperoxidase-labelled secondary antibody directed againstmouse immunoglobulin G (Fig. 4a). Obviously, the LA-

ICP-MS-based method using DOTA-chelated europiumattains similar although slightly lower sensitivity as theluminescence-based procedure. The latter, however, is notsuitable for multiplexing as will be discussed later.

Fig. 3 127I signal intensities as a function of amounts of microsomalprotein applied. The intensities from Fig. 2 were integrated and usedfor the plot. There is a linear relationship between the iodine signaland the amounts of antigen applied in SDS-PAGE separation

Fig. 2 Blot of SDS-PAGE sepa-rated iodine-labelled monoclonalCYP2E1 antibody under reducingconditions analysed by LA-ICP-MS for 127I+. The amount ofantibody applied (micrograms) isindicated. The signal intensityscale is on the right

Fig. 4 Detection of SDS-PAGE separated and blotted CYP1A1 fromrat liver microsomes of 3-methylcholanthrene-treated rats. Amount ofmicrosomal protein applied 0.1–to 5 μg as indicated. a Primaryantibody, anti-CYP1A1; secondary antibody, goat antimouse immu-noglobulin G, peroxidase coupled; detection, luminescence. b Primaryantibody, anti-CYP1A1, europium-labelled; detection, LA-ICP-MS,153Eu+

Detection of electrophoretically separated cytochromes P450 1141

Detection of blotted CYP2E1 with an iodine-labelledantibody

For detection of CYP2E1, we developed a specific mono-clonal mouse antibody. It was labelled by iodine using acompletely different chemistry than for DOTA-lanthanidelabelling of the CYP1A1 antibody. Differential labelling ofthe antibodies was done for discrimination in multiplexinganalyses.

The experiment performed is similar to that described forCYP1A1 detection; however, we used liver microsomes of ratstreated with isoniazid, which is known to increase CYP2E1levels [46]. Applied protein amounts were between 0.125 and5.0 μg. The 127I+ signal obtained by LA-ICP-MS showeddefined bands at about 50 kDa (Fig. 5) and a clear dependenceon the protein amount. A signal could be identified down tothe lowest amount of protein applied (0.125 μg).

A parallel analysis of the samples by means of asecondary antibody and luminescence detection showedthat both methods have comparable sensitivity. The de-tection limit for CYP2E1 is close to 0.2 μg appliedmicrosomal protein. A calculation shows that this corre-sponds to about 70 fmol CYP2E1. The calculation is basedon the specific CYP content of the microsomes which

amounts to about 1 nmol total CYP per milligrammicrosomal protein and the fact that CYP2E1 constitutesabout 30% of total CYPs in liver microsomes of isoniazide-treated rats [45].

Aspects of CYP quantitation by LA-ICP-MS

Routinely, the two microsome preparations described in theprevious sections serve as standards in our laboratory forsemiquantitative CYP1A1 and CYP2E1 measurements viaimmunoblots and luminescence detection [9, 10]. Usually,two samples of a preparation with different protein amountsare run on each gel for calibration together with the samplesto be analysed either for CYP1A1 or CYP2E1. For the timebeing, we will adopt this procedure also for analyses byICP-MS. A calibration curve for CYP2E1 obtained withuse of our standard microsomes is shown in Fig. 5. Thesemicrosomes were prepared in large amounts from livers ofrats treated with the CYP2E1 inducer isoniazid. The finalgoal, however, is to produce standards for absolutequantitation, i.e. purified CYP enzymes. These can be usedeither directly as standard proteins or to determine thespecific CYP content of the standard microsomal prepara-tion which is then further used for calibration. The latter

Fig. 5 Detection of CYP2E1 in liver microsomes of isoniazid-treated ratswith LA-ICP-MS by means of an iodinated monoclonal antibody specificfor CYP2E1. Microsomal proteins in the amounts stated (micrograms)were separated by SDS-PAGE and subsequently transferred to anitrocellulose membrane by electroblotting. b The membrane was

incubated with an iodine-labelled monoclonal antibody with specificityfor rat CYP2E1 and 127I+ was quantified by LA-ICP-MS. a Blotmembrane with the same microsomal samples and amounts as in b butwith luminescence-based detection of microsomal CYP2E1 by means ofa peroxidase-coupled secondary antibody

1142 P.H. Roos et al.

constitutes a cost-saving alternative compared with exclu-sive application of purified enzymes. Corresponding experi-ments and method developments are currently beingperformed. For both the semiquantitative and the quantita-tive analyses via protein standards, the labelling degree ofthe antibodies has not to be taken into account.

Another approach for CYP protein quantitation via LA-ICP-MS could be via absolute quantitation of the labelelements, in our case europium and iodine. This requires (1)calibration with known amounts of label, (2) knowledge ofthe number of labels per antibody and (3) the antibody-antigen binding stoichiometry. The latter is theoretically 1:1

for monoclonal antibodies and monomer CYPs assuminghigh affinity of the antibody for the antigen. The method ismore laborious than the aforementioned one and includesan assumption of antigen-antibody stoichiometry, so weprefer to use CYP standards for calibration.

Multiplexing

For multiplexing, allowing for concomitant detection ofCYP1A1 and CYP2E1, we used three different microsomalpreparations: (1) liver microsomes of 3-methylcholan-threne-treated rats with a known high content of CYP1A1

Fig. 6 Concomitant detection of CYP1A1 (a) and CYP2E1 (c) with aeuropium-labelled and an iodine-labelled antibody, respectively. Blotmembranes were scanned and analysed by LA-ICP-MS for 127I+ and153Eu+. b The 5 and 10 μg samples as in c but developed via the

CYP2E1 antibody by luminescence detection on the blot membrane.The samples were liver microsomes of 3-methylcholanthrene-treatedrats. The amounts of protein applied (micrograms) are indicated

Detection of electrophoretically separated cytochromes P450 1143

and a moderate CYP2E1 level, (2) liver microsomes ofisoniazid-treated rats with increased CYP2E1 and at bestbackground CYP1A1 levels and (3) duodenal microsomesof minipigs treated orally with a mixture of PAHs. TheCYP content of minipig duodenal cells is much lower thanthat of liver cells [10]. However, PAHs have been shown toinduce CYP1A1-dependent enzymatic activity in cells ofthe minipig duodenum [10]. Little is known on the presence

and the level of CYP2E1 in these cells. Thus, by selectionof the different microsomal preparations, the LA-ICP-MSmethod is confronted with easy and more difficult tasksallowing a first evaluation of the method.

Liver microsomes of 3-methylcholanthrene-treated rats

Microsomes subjected to SDS-PAGE in amounts between 0.5and 10 μg were PAGE-separated and blotted. The blot wasconcomitantly developed with the europium-labelledCYP1A1 and the iodine-labelled CYP2E1 antibody. Thereadout from LA-ICP-MS shows specific signals for CYP1A1and CYP2E1 detected in the same lane (Fig. 6a,c). However,additional iodine-positive bands corresponding to apparentmolecular masses greater than 50 kDa are observed. Theorigin of these bands is not clear but we assume that theyrepresent ubiquitinated CYP2E1 protein still recognized bythe antibody. Ubiquitin is a small protein attached by ligasesto other proteins and is involved in the process of proteindegradation. In fact, ubiquitination has been observed forCYP2E1 [47, 48]. Because the bands detected by LA-ICP-MS are also seen by luminescence-based CYP2E1 detection(Fig. 6b), they do not appear to be artefacts of the LA-ICP-MS method. Figure 7 shows that CYP1A1 and CYP2E1 arediscriminated by the element labels and also by electropho-retic mobility. Rat CYP1A1 and CYP2E1 have molecularmasses of 59,393 and 56,626 Da, respectively, and thusdiffer by about 2.8 kDa, but are very well separated by about5 mm in terms of scanning distance.

Fig. 7 Superimposed line scans for 153Eu+ and 127I+ of the 10 μgsample of Fig. 6 corresponding to CYP1A1 and CYP2E1, respectively.The main protein peaks are electrophoretically separated. A CYP2E1-derived rider peak overlaps with the CYP1A1 peak clearly and they aredistinguished by the different labels. 3MC 3-methylcholanthrene

Fig. 8 Concomitant detection of CYP1A1 (a) and CYP2E1 (b) with aeuropium-labelled and an iodine-labelled antibody, respectively. Blotmembranes were scanned and analysed by LA-ICP-MS for 127I+ and

153Eu+. The samples were liver microsomes of isoniazid-treated rats.The amounts of protein applied are indicated

1144 P.H. Roos et al.

Liver microsomes of isoniazid-treated rats

Microsomes were subjected to SDS-PAGE in amounts of 0.5,1.0 and 2.0 μg. While CYP2E1 is clearly detectable on theblot membrane by the 127I+ signal down to 1 μg appliedmicrosomal protein, a CYP1A1-associated 153Eu+ signal isnot or is only slightly seen even at 2 μg protein (Fig. 8). Theoverlay of the line scans corresponding to the 2 μg sampleclarifies this fact (Fig. 9). In conclusion, the results confirmwhat is known about the CYP expression profile in livermicrosomes of isoniazid-treated rats. It is noted that CYP2E1makes up about 1.6% of total liver microsomal protein in

these microsomes [45]. Thus, 1 μg protein applied in theanalysis corresponds to about 160 fmol CYP2E1.

Duodenal microsomes of minipigs orally exposed to PAHs

Because CYP levels are low in microsomes of the gastroin-testinal tract, amounts of 10 and 20 μg protein were used forthe analyses. In contrast to liver microsomes of isoniazid-treated rats, CYP1A1 is clearly expressed in duodenalminipig microsomes after PAH exposure (Fig. 10). CYP2E1is also detectable but only at low levels (Fig. 10). Two blotmembranes developed by means of the primary antibodiesand a peroxidase-coupled secondary antibody were used forseparate detection of CYP1A1 and CYP2E1 by lumines-cence. The results shown in Fig. 10a and b are inaccordance with the LA-ICP-MS data. Studies on CYPsin minipigs have been largely confined to the liver [49, 50].Furthermore, there have only been a few studies oninduction of CYP enzymes in minipigs by xenobiotics.Oral intake of PAHs leads to induction of CYP1A1 in liver,kidneys, lung and also duodenum of minipigs [10]. Thepresence of CYP2E1 protein in minipig duodenal cells isshown here for the first time. Besides the usual CYP bandat about 50 kDa, a further band corresponding to lowermolecular mass is detected. Its identity is not clear, but apossible explanation is that it constitutes a CYP2E1 variantbased on alternative splicing as was proven by Bauer et al.[51] for human CYP2E1. They detected CYP2E1 mRNAvariants missing exon 2, exons 2,3 and exons 2,4,6.Generally, little or nothing is known on CYP2E1 splicevariants in humans and minipigs, respectively. We do not

Fig. 9 Superimposed line scans for 153Eu+ and 127I+ of the 10 μgsample of Fig. 8 corresponding to CYP1A1 and CYP2E1, respectively

Fig. 10 Concomitant detection ofCYP1A1 (c) and CYP2E1 (d)with a europium-labelled and aniodine-labelled antibody, respec-tively. The samples were duode-nal microsomes of minipigs orallyexposed to polycyclic aromatichydrocarbons (the amounts ofprotein applied are indicated).Blot membranes were scannedand analysed by LA-ICP-MS for127I+ and 153Eu+. In parallel, asecond set of blots was treatedwith a primary anti-CYP1A1 an-tibody (a) or anti-CYP2E1 anti-body (b) and a secondaryperoxidase-coupled goat anti-mouse immunoglobulin G anti-body. The bands were visualizedby chemiluminescence detection

Detection of electrophoretically separated cytochromes P450 1145

consider the lower molecular mass band to be a proteolyticfragment because we used a cocktail of protease inhibitorsfor preparation of duodenal microsomes. The unidentifiedband is also visible on the luminescence-based blot but israther faint (not shown), i.e. it is detectable with highersensitivity by LA-ICP-MS than by the luminescencemethod. This also applies to the higher molecular massCYP2E1 bands seen in Fig. 6b and c. Currently, we areinvestigating this phenomenon.

Conclusion

A novel approach to detect and identify different CYPssimultaneously in a single step by means of co-incubationwith iodinated and lanthanide-labelled antibodies has beendemonstrated. For the analyses, only ‘primary’ antibodiesare necessary. In contrast, a primary antibody is combinedwith a peroxidase or otherwise labelled secondary antibodyin conventional blot analyses. Differential antibody label-ling with elements such as lanthanides or iodine allows amultiplexing approach by the LA-ICP-MS method, whilethis is usually not possible by means of secondaryantibodies, particularly in the case of overlapping bands.As shown here, the sensitivity of the luminescence methodis comparable to that of the LA-ICP-MS method, while thelatter excels because of the multiplexing aspect. Our currentinvestigations aim to measure multiple CYPs, includingcomponents of their regulation pathways, at the same time.In the next steps of method development, we intend toincrease the number of parameters to be analysed byinclusion of further differentially labelled antibodies andto prepare standards for absolute quantification. Furtherimprovements are possible which can yield significantsignal amplification if more than one label per linkingmolecule is used [52].

Acknowledgements We are indebted to Ingeborg Bichbäumer forsample preparation. The work of ISAS and of IfADo was financiallysupported by the Bundesministerium für Bildung und Forschung andby the Ministerium für Innovation, Wissenschaft, Forschung undTechnologie des Landes Nordrhein-Westfalen.

References

1. Omura T, Sato R (1964) J Biol Chem 239:2370–23782. Nelson D (2008) Cytochrome P450 homepage. http://drnelson.

utmem.edu/CytochromeP450.html3. Gelboin HV, Krausz K (2006) J Clin Pharmacol 46:353–3724. Hankinson O, Brooks BA, Weir-Brown KI, Hoffman EC,

Johnson BS, Nanthur J, Reyes H, Watson AJ (1991) Biochimie73:61–66

5. Riddick DS, Huang Y, Harper PA, Okey AB (1994) J Biol Chem269:12118–12128

6. Omiecinski CJ, Redlich CA, Costa P (1990) Cancer Res 50:4315–4321

7. Roos PH, van Afferden M, Strotkamp D, Pfeifer F, Hanstein WG(1996) Arch Environ Contam Toxicol 30:107–113

8. Kondraganti SR, Muthiah K, Jiang W, Barrios R, Moorthy B(2005) Chem Res Toxicol 18:1634–1641

9. Roos PH (2002) Arch Toxicol 76:75–8210. Roos PH, Tschirbs S, Welge P, Hack A, Theegarten D, Mogilevski

G, Wilhelm M (2002) Arch Toxicol 76:326–33411. Roos PH, Bolt HM (2008) In: Ioannides C (ed) Cytochromes

P450-role in the metabolism and toxicity of drugs and otherxenobiotics. Royal Society of Society, London, (in press)

12. Kawajiri K, Fujii-Kuriyama Y (1991) Jpn J Cancer Res 82:1325–1335

13. Ioannides C, Parke DV (1993) Drug Metab Rev 25:485–50114. Nebert DW, Dalton TP, Okey AB, Gonzalez FJ (2004) J Biol

Chem 279:23847–2385015. Rivera SP, Saarikoski ST, Hankinson O (2002) Mol Pharmacol

61:255–25916. Nebert DW, Puga A, Vasiliou V (1993) Ann N Y Acad Sci

685:624–64017. Bolt HM, Roos PH, Thier R (2003) Int Arch Occup Environ

Health 76:174–18518. Yamazaki H, Inui Y, Yun CH, Guengerich FP, Shimada T (1992)

Carcinogenesis 13:1789–179419. Thomas PE, Bandiera S, Maines SL, Ryan DE, Levin W (1987)

Biochemistry 26:2280–228920. Palakodety RB, Clejan LA, Krikun G, Feierman DE, Cederbaum

AI (1988) J Biol Chem 263:878–88421. Ingelman-Sundberg M, Ronis MJ, Lindros KO, Eliasson E,

Zhukov A (1994) Alcohol Alcohol Suppl 2:131–13922. Koop DR, Tierney DJ (1990) Bioessays 12:429–43523. Gonzalez FJ, Ueno T, Umeno M, Song BJ, Veech RL, Gelboin

HV (1991) Alcohol Alcohol Suppl 1:97–10124. Vieira I, Sonnier M, Cresteil T (1996) Eur J Biochem 238:476–

48325. Hu S, Zhang S, Hu Z, Xing Z, Zhang X (2007) Anal Chem

79:923–92926. Whetstone PA, Butlin NG, Corneillie TM, Meares CF (2004)

Bioconjug Chem 15:3–627. Ahrends R, Pieper S, Kühn A, Weisshoff H, Hamester M,

Lindemann T, Scheler C, Lehmann K, Taubner K, LinscheidMW (2007) Mol Cell Proteomics 6:1907–1916

28. Liu H, Zhang Y, Wang J, Wang D, Zhou C, Cai Y, Qian X (2006)Anal Chem 78:6614–6621

29. Moreau J, Pierrard JC, Rimbault J, Guillon E, Port M, AplincourtM (2007) Dalton Trans 28:1611–1620

30. Sosabowski JK, Mather SJ (2006) Nat Protoc 1:972–97631. Ornatsky O, Baranov VI, Bandura DR, Tanner SD, Dick J (2006)

J Immunol Methods 308:68–7632. Tanner SD, Ornatsky O, Bandura DR, Baranov VI (2007)

Spectrochim Acta B 62:188–19533. Razumienko E, Ornatsky O, Kinach R, Milyavsky M, Lechman E,

Baranov V, Winnik MA, Tanner SD (2008) J Immunol Methods336:56–63

34. Roos PH, Hartmann H-J, Schlepper-Schäfer J, Kolb H, Kolb-Bachofen V (1985) Biochim Biophys Acta 847:115–121

35. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) J BiolChem 193:265–275

36. Sakaki T, Soga A, Yabusaki Y, Ohkawa H (1984) J Biochem(Tokyo) 96:117–126

37. Ryan DE, Ramanathan L, Iida S, Thomas PE, Haniu M, ShivelyJE, Lieber CS, Levin W (1985) J Biol Chem 260:6385–6393

1146 P.H. Roos et al.

38. Roos PH (1999) Xenobiotica 29:629–64039. Kearney JF, Radbruch A, Liesegang B, Rajewsky K (1979) J

Immunol 123:1548–155040. Falkenberg FW, Weichert H, Krane M, Bartels I, Palme M, Nagels

HO, Fiebig H (1995) J Immunol Methods 179:13–2941. Guengerich FP (1977) J Biol Chem 252:3970–397942. Garijo Anorbe M, Messerschmidt J, Feldmann I, Jakubowski N

(2007) J Anal At Spectrom 22:917–92443. Venkatachalam A, Koehler CU, Feldmann I, Lampen P, Manz A,

Roos PH, Jakubowski N (2007) J Anal At Spectrom 22:1023–103244. Feldmann I, Koehler CU, Roos PH, Jakubowski N (2006) J Anal

At Spectrom 21:1006–101545. Roos PH (1996) J Chromatogr B 684:107–131

46. Ryan DE, Ramanathan L, Iida S, Thomas PE, Haniu M, ShivelyJE, Lieber CS, Levin W (1985) J Biol Chem 260:6385–6393

47. Roberts BJ, Song BJ, Soh Y, Park SS, Shoaf SE (1995) J BiolChem 270:29632–29635

48. Roberts BJ (1997) J Biol Chem 272:9771–977849. Anzenbacher P, Soucek P, Anzenbacherova E, Gut I, Hruby K,

Svoboda Z, Kvetina J (1998) Drug Metab Dispos 26:56–5950. Skaanild MT (2006) Curr Pharm Des 12:1421–142751. Bauer M, Herbarth O, Aust G, Graebsch C (2005) Mol Cell

Biochem 280:201–20752. Lou X, Zhang G, Herrera I, Kinach R, Ornatsky O, Baranov V,

Nitz M, Winnik MA (2007) Angew Chem Int Ed 46:6111–6114

Detection of electrophoretically separated cytochromes P450 1147