characteristic

Stokes shift

EmissionbandLJfetime

conventionalfluerophorea

Small (28 nm forfluorescein)

WideShort (<100 ns)

Significant

Significant

DifficultSevereSignificant

Excellent-iO- mol/L

Lanthanlde

Large (290-300nm for Eu3)

NarrowVery long

(1-1000 L5)

Negligible-nonexistent

Negligible

SimpleNonexistentNonexistent

Good_.10_13 mol/L

Department of Clinical Biochemistry, Toronto Western Hospi-tal, 399 Bathurst St., Toronto,Ontario M5T2S8, Canada (addressfor correspondence), and Department of Clinical Biochemistry,University of Toronto,Toronto,Ontario M50 1L5.

Received April 16, 1991; accepted July 2, 1991.

CLIN.CHEM.37/9, 1486-1491(1991)

1488 CLINICALCHEMISTRY,Vol.37, No.9, 1991

Multiple Labeling and Time-Resolvable FluorophoresEleftherios P. Diamandis

A new time-resolved fluorescence immunoassay systeminvolving use of the europium chelate of 4,7-bis(chloro-sulfophenyl)-1, 1O-phenanthroline-2,9-dicarboxylicacid aslabel is reviewed. This stable chelate by itself is not veryfluorescent but, used in multiple labeling strategies, im-proves the achievable detection limits. By using multiplelabeling, streptavidin tailing, and Eu3 activation, one cancreate a very stable, easy-to-use reagent that is suitablefor devising highly sensitive immunoassays and otherbiotechnological assays. This reagent, a streptavidin-based macromolecular complex, is able to detect-300 000 molecules (-0.5 amol) of aipha-fetoprotein in amodel noncompetitive immunoassay.

AddItIonalKeyphrase.: fluoroimmunoassay . sfreptavidin

Conventional fluorophores have been used for manyyears in various biotechnology applications, includingimmunoassay (1, 2). Currently, these fluorophores areroutinely used as labels in several automated homoge-neous imniunoassay procedures applicable to analytespresent in serum at micromolar or nanomolar concen-trations. No system is commercially available that usessuch fluors for immunoassay procedures having sensi-

tivity in the picomolar or subpicomolar range, becausethe detection sensitivity of conventional fluors is se-verely compromised by background signals. Such back-ground comes from light scatter, fluorescence of endog-enous compounds in the biological sample (e.g., pro-teins, biirubin, etc.), fluorescence of the cuvettes, optics,solvents, solid phases used in immunoassay, etc. Thevery small Stokes shifts and wide emission bands ofsuch fluors (i.e., 28 nm Stokes shift for fluorescein) is animportant limitation for several different reasons, in-cluding the following: (a) excitation radiation cannot becompletely excluded from reaching the detector becauseof scattering; (b) in minimizing the scattering effect,narrow bandpass emission filters are used, which ex-clude a significant portion of the emitted radiation; and(c) the wide emission band is likely to overlap with

emission bands of interfering compounds.Fluorescent chelates of europium, terbium, and samar-

ium are potential alternatives to isotopic labels, conven-tional fluorophores, and other nonisotopic labels for immu-nological and other assays. The main advantages of thesefluorescent lanthanide chelates are shown in Table 1. Bycombining the advantages listed, background fluorescence

can be virtually eliminated (Figure 1). Thus, a pulsed-

light time-resolved fluorometric measurement of a eu-ropium chelate is at least 103-fold more sensitive than is ameasurement based on conventional fluorometry.

Lanthanide Chelates for ImmunologIcal Assays

The first proposal of lanthanide chelates as labels foriinmunoassay applications was put forward by Wieder(3) and Leif et al. (4). A lanthanide chelate can be usedas an immunological label in a simple assay configura-tion (Figure 2), which is identical to a configurationused for immunoradiometric assays and other noniso-topic assays. Similar immunocomplexes can be formedwith competitive-type immunoassays. However, thissimple assay configuration has been very difficult to usesuccessfully, for the following reasons:

#{149}Europium chelates absorb radiation characteristicof the ligand and emit radiation characteristic of themetal ion. The fluorescence quantum yield of the chelatedepends on the efficiency of the energy transfer processfrom the ligand to the ion. If the energy transfer isinefficient, the chelate is not very fluorescent and sen-sitivity is relatively poor. The most fluorescent com-plexes of europium are complexes of the typeEu(NTA)3(TOPO)2, where NTA = naphthoyltrifluoro-acetone and TOPO = trioctylphosphine oxide. The 1:1complexes between Eu3 and various ligands are notvery fluorescent.

#{149}Apart from the requirement that the chelata be fluo-rescent, it must also possess very high stability; otherwise,

during the washing process necessary for heterogeneousiinmunoassays, Eu3 would leak into the solution.

Table 1. Advantages of Lanthanide Chelates InComparison wIth ConventionalFluorophores

Overlap of excitation-emission spectra

Overlap with serumautofluorescence

PulseexcitationaUghtscatterBackground

fluorescenc&’QuantumyieldDetectionlimit

#{149}Usedwith microsecond time-reso Wed fluorometty.#{176}Optics,solid phases, solvents, impurities.

0lJ

cs-

CLINICALCHEMISTRY,Vol.37, No.9, 1991 1487

.

..:

‘#{149}‘ : #{149}:#{149}:,#{149}#{149},i#{149}1.. #{149} .S

#{149} .1

Conventional Time-ResolvedFluorescence Fluorescence: 20 ns 200 jjs

ins after flash startmeasurementexcitation

Fig. 1. SchematIc diagram showing comparison of a conventionalfluorometricmeasurementand a time-resolved fluorometnc mea-surementShortly aftera pulsed excitation, both short-lived and long-lived fluors emit, butat 20 ns most of the short-livedfluorescence has dissipated. At 200 is, onlythe long-lived fluors emit, which can be measured underconditionsof zerobackground with very high sensitivity. This Figure has no quantitative meaning

Fig. 2. “Two-site” (sandwich) assay for an analyte: one of theantibodIes (capture) Is immobilizedon a solid support; the other(detection) Is labeled witha europiumchelateThe tigand binds Eu3 strongly so that no leakage occurs during extensivewashing. Ideally, the ligand absorbs excitation energy and transfers Itto Eu3,which Is, In turn, excited and subsequently emits long-lived fluorescence

Although forming very highly fluorescent complexesof Eu3 with organic ligands (with long decay times,suitable for microsecond time-resolved fluorometry) iseasy, these complexes are not soluble and stable enoughto be linked to detection antibodies. On the other hand,the highly stable 1:1 complexes of Eu3 and variousammopolycarboxylic acids can be used successfully tolabel antibodies with Eu3 but are not very fluorescentbecause the energy-transfer process is not optimal withsuch chelates. Thus, the first time-resolved fluorometriciinmunoassay system described and commercialized isbased on the following two steps:

#{149}A complex (as shown in Figure 2) is formed by usingdetection antibodies that carry Eu3 through a nonflu-orescent chelate with an aminopolycarboxylic acid de-rivative covalently linked to the antibody.

#{149}To measure Eu3, the derivative-antibody conju-gate is released in solution at low pH and is recomplexedwith appropriate organic ligands to form a highly fluo-rescent complex in solution.

The system, known as DELFIA (Pharmacia-LKB,

Bromma, Sweden) has excellent sensitivity, broad dy-namic range, and reagent stability. The excellent sen-sitivity is due to the fact that, during the immunoreac-tion, Eu3 is carried on the antibody tightly bound to astrong chelator without any leakage. Also, during mea-surement, Eu3 is optimally complexed by organic hg-ands to form a highly fluorescent complex. However,this system has been criticized for three reasons: (a) Anextra step is needed to extract Eu3 from the solid phasebefore measurement; (b) more importantly, the final

measurement is based on Eu3 quantification, usuallyat 10- 12 1O_13 mol/L, but because Eu3 is a naturalelement present in dust, skin, etc., the possibility ofcontamination, especially of the reagents containing theeuropium chelators, presents difficulties; (c) because thefinal measurement is based on extraction, the system isnot suitable for other biospecific assays, e.g., Southern,Northern, or Western blotting, DNA sequencing, orimmunohistochemistry and flow cytometry.

Immunoassays wIth BCPDA-Labeled DetectIon

AntIbodIes

About six years ago, we started research towards thesynthesis of organic europium chelators that would com-bine good complex stability and energy transfer so thatwe could achieve assays based on the design of Figure 2without the need for extraction of Eu3. Such efforts havealso been undertaken by other groups; isolated reportsare described by Soini et al. (5). The best chelator iden-tilled by our group, BCPDA, was synthesized as shown inFigure 3 (6).1 BCPDA has the following structural fea-tures: a polyaromatic phenanthroline ring, which canabsorb radiation subsequently used to internally excitecoordinated Eu3, and a Eu3-binding site consisting oftwo heteroaromatic nitrogens and two carboxylic acids.Eu3 can form either 1:1 or 1:2 complexes with BCPDA;

the Eu(BCPDA)2 complexes are more fluorescent than the1:1 complexes (7, 8). BCPDA also possesses two sulfonyl-chloride groups, which are used to link the molecule toproteins through e-amino groups of lysines or to otherprimary amine-containing compounds (7).

Earlier reports by our group established optimal pro-cedures for labeling proteins with BCPDA (7). In general,we found it easy to derivatize all amino groups of manydifferent proteins; however, for biologically active mac-romolecules, e.g., antibodies, avidin, streptavidin, ex-

‘Nonstandard abbreviations: BCPDA, 4,7-bia(chloroeulfophe-nyl)-1,10-phenanthroline-2,9.dicarboxylic acid; SBMC, streptavi-din-based macromolecular complex; SA, atreptavidin and TG,bovine thyroglobulin.

iv -

Fig. 3. Synthesis of 4,7-bls(chlorosulfophenyl)-1 ,10-phenanthroline-2,9-dicarboxylic acid (BCPDA) (IV) from the commercially availablecompound bathocuprolne (I)Adapted from reference 6 by pemlieslon

STEP I:

+

HQ3S o 0

0

STEP 2

- rU2$-..NH2 “-4r61 o

(s-s) +

H2H>’HH2 _rr ,,

DSA

BCPDA

Su1IoSMCC

*

*

KNH4R.teOj) !!!!.i + OTT0 #{149} *

su

*4 l-*HS.r’SH

*

-:o:

STEPS

ITSignal amplification15-18-fold

Signal amplification40-fold

Fig. 5. Conceptual view of an assay withdirectly labeled BCPDA-detection antibody (left) and antibody covalently linked to BCPOA-labeled bovine serum albumin (C, right)

1488 CLINICALCHEMISTRY,Vol.37, No.9, 1991

tensive labeling results in partial or full loss of biologi-cal activity. Thus, optimal labeling is mandatory, notjust desirable (7). Below, I describe how BCPDA was usedas a label to devise highly sensitive time-resolved fluo-rescence immunoassays.

Labeling

Polyclonal and monoclonal anti-cortisol antibodiesand other monoclonal detection antibodies were success-fully labeled with BCPDA with a load of 12-18 BCPDAmolecules per antibody molecule without significantloss of the antibody’s binding activity. Additionally,second antibodies (e.g., goat anti-mouse IgG) were alsolabeled successfully with BCPDA at the above ratios.These antibodies were used successfully for both com-petitive-type and noncompetitive-type assays (e.g., inthe configuration shown in Figure 2), with the fluores-

cent immunocomplex measured directly on the solidphase. Although it was clearly established that in suchassays, with BCPDA as label, the immunocomplex doesnot lose Eu3 during repeated washing and is fluores-cent so that there was no need for Eu3 extraction(unpublished data), the configuration was not highlysensitive. We attributed the low sensitivity partly to thefact that only 1:1 complexes of Eu3 with BCPDA couldform because of the presence of excess Eu3.

The issue of low sensitivity was addressed in a seriesof efforts to amplifr the above system so that highlysensitive assays could be achieved.

Antibody Labeling

Antibodies possess many s-amino groups (>50 permolecule) but derivatization of >18 per molecule withBCPDA inactivates the antibody. We decided to increasethe antibody load by covalently linking it with a carrierprotein that is already highly labeled with BCPDA. Theconjugation chemistry used is diagrammatically shownin Figure 4. For the carrier protein, we used bovineserum albumin, which could carry as many as 40 BCPDAmolecules (Figure 5). The antibody conjugate was usedsuccessfully in a competitive-type assay of cortisol witha signal amplification about double that of the directlylabeled antibodies (9).

Biotin-Streptavidin System

This versatile system has recently been reviewed (10).In most immunoassay applications, biotin is introducedinto detection antibodies at a load of 10-20 biotins perantibody, by reacting the s-amino groups of lysines withan N-hydroxysuccinimide ester of biotin in a very sim-ple one-step reaction. Streptavidin (SA), labeled with areporter molecule, is added in a separate step to bind tobiotin and thus reveal the presence of the imniunocom-plex. A final immunocomplex on the solid phase for anaipha-fetoprotein assay is shown in Figure 6 (left). Thebiotin-streptavidin system introduces amplification inan immunoassay because (a) at least 10-15 biotins couldbe available to bind labeled strept.avidins and (b)streptavidin can carry as many as 14 BCPDA moleculeswithout any loss of its biotin-binding activity (7). Thus,

Fig. 4. Procedure for the preparationof antibody-bovine serumalbumin(BSA)-ecPDAconjugatesIn step 1, antIbodyis reacted with sulfosucclnlmidyl 4-(NmaIeimIdomethyl)cy.clohexane-1-carboxylate (sulfo-SMCC) to introduce maleimide groups. Instep2, BSA is coupled to 8CPDA. Instep 3,labeledBSA isreduced by dlthlothreltol(OTT)to create free SH groups. Instep 4, the derivatized antibody is coupledto reduced BSA. n, x and y represent the numberof total amino groups,derivatized amino groups, and conjugated amino groups, respectively. Re-printedfrom reference 9by permission

the total amplification can be as high as 14 x 15 =

210-fold in comparison with only 15- to 18-fold achievedby using directly labeled antibodies. An extra 16-foldamplification in signal has been observed in a modelalpha-fetoprotein assay when SA(BCPDA)14 and biotiny-lated detection antibodies are used instead of BCPDA-labeled detection antibodies (11).

Streptavidin Labeling

In our previous studies of labeling various proteinswith BCPDA (7), we noticed that we could easily incorpo-rate into the proteins as many or, in some instances,

(5*-NH2 ) (To.Nu2)

NH2

+ +

0 0[ ICH2_h N -OC -CH2- S - C - CU3

0 0 SATA

SLLFO.SMCC

pH 9.1

Signal amplllicatlon1500-2250told(1010 15 x 150)

pHi

5 A-NH...

pH6.2N2

-4

NH2OH

0 0Ii

T0.NH-C-C142-S-C -CU3

NH2

pH9.1

STREPTAVIOIN

CLINICALCHEMISTRY,Vol.37, No.9, 1991 1489

0,,,,

Signalamplification140-210-told11010 15 x 14)

Fig. 6. Schematic of an assay invoMng BCPDA-labeled streptavidin(SA(BcpDA)1J (lefi) and streptavidincovalently linked to BCPDA-labeled thyroglobulin [SA(TG)(BcPoA)1,j (right)Biotin (arrovd)Is shown bound to streptavidin. -0 Indicates the Eu3-scPDA

complex

more BCPDA molecules than the number of s-aminogroups. When we checked these proteins for fluorescencein the preseilce of excess Eu3, we noticed two importantfacts. First, no fluorescence quenching was present,although this fluorescence quenching occurs to a signif-icant degree with polyfluoresceinated proteins or nu-cleic acids (12, 13). Fluorescence quenching with multi-ple fluorescence labeling occurs because of the signifi-

cant overlap between excitation and emission spectra;thus, it is possible for one fluor to emit fluorescence andan adjacent fluor to absorb it. The closer the fluors, thegreater the quenching. With fluorescent Eu3 chelates(and probably also with fluorescent Th3 and Sm3chelates), such quenching is not expected and was notobserved because there is absolutely no overlap betweenthe excitation and emission spectra. Second, the fluores-cence found was not just proportional to the degree ofmultiple labeling with BCPDA but was about three- tosixfold greater. This “bonus effect” was considered to berelated to environmental effects of the various aminoacids around the Eu3 and BCPDA.

Bovine thyroglobulin (TG) is a high-Mr protein (660 000Da) carrying -150 s-amino groups (7). We were able tointroduce --175 BCPDA labels per molecule of TG andobtain fluorescence equivalent to that obtained by -900BCPDA residues, with excess Eu3. We decided to create anew biotin-binding protein that could carry not just 14BCPDA molecules but -150, instead. We covalently linkedstreptavidin to BCPDA-labeled TG with an elegant conju-gation chemistry shown in Figure 7 and further describedelsewhere (11). The new streptavidin reagent, representedby the formula SA(TGXBCPDA)ise, gave signal amplifica-tion and detection limit improvement by about fivefoldthat of SA(BcPDA)14 (directly labeled streptavidin). Obvi-ously, some steric hindrance effects of streptavidin bindingdo occur with this newer reagent, but the benefits arequite significant (Figure 6, right).

Streptavidin-Based Macromolecular Complex (SBMC)

A serendipitous observation during accelerated stabil-ity studies with the reagent SA(TG)(BcPDA)1 led to thedevelopment of an SBMC, which now is our reagent of

0 0I I

T0-NH-C-CH2-S-C 0H3NH

ec,oA

5*-NH - -- CH2_PJ’

DC P0*

Fig.7. Schematic representation of the conjugation reaction betweenstreptavidin (SA) and BCPDA-labeledthyroglobulin (TG)StreptavldinIsdeilvatized withsulfo-SMCC (see Fig. 4) to introduce maleimidegroups. TG is first thiolated with succinlmldylacetylthioacetate (SATA) (-15thiols/TO) and then extensively labeled withBCPOA (-150 ecPodTG). Thetwoderivatives are then reacted under nitrogen and Inthepresence of hydroxyl-amIne, which liberates free thiols. The conjugate Is SA(TG)(BceoA),. Re-printed from reference 11 by permission

choice for both immunological and nucleic acid hybrid-ization assays (14). During the conjugation reactionshown in Figure 7, we routinely use a fivefold molarexcess of thiolated-BCPDA-labeled TG over streptavidinderivatized with succinimidyl 4-(N-maleimidomethyl)-cyclohexane-1-carboxylate, to promote quantitative con-jugation of streptavidin. Routinely, we do not remove theexcess of thiolated-BCPDA-labeled TG because we havenot observed any significant increase in the backgroundsignal with its presence. However, when we added Eu3’to the final reaction mixture after conjugation, whichcontained SAcrG)(BcPDA)1 and TG(BcPoA)1, and in-cubated at 50-55 #{176}Cfor a few hours, we observed that anew, more sensitive reagent is formed under controlledreaction conditions, especially when Eu3 concentrationis carefully controlled. After extensive studies (14), weproposed that under such conditions, a SBMC is formed(Figure 8), consisting of (a) SA(TG)(BcPDA)1, (b) TG(B-CPDA)1, and (c) Eu3 at a molar ratio of about 1/3.3/480.We postulated that this SBMC is formed by Eu3 actingas a linker between components a and b. The SBMC isvery stable and does not disintegrate during the immu-noassay procedure or during extensive washings. Wefound that no Eu3 leakage occurs even if the tracer

diluent does not contain Eu3. Moreover, we found thatthe SBMC is stable for years when kept as a stocksolution containing 15 mg of streptavidin per liter. Theoperational structure of the SBMC is SA(TG)3.3-(BCPDA) with a Mr of -(2.5-3) x iOn. The theoreticalamplifications with various reagents described here areshown in Table 2.

Sireptavidinbasedmacromolecularcomplex(SBMC)

Fig. 8. Schematic representationof the proposed mechanism offormationof the streptavidin-basedmacromolecularcomplex;corn-plexed with blotin (not shown), the system gives a signal amplifica-tionof 4500- to 6750-fold-0 = BCPOA. The Figure has no quantitative meaning. For more details,seetext and reference 14

Ab(BCPDA)118Ab(BSA)(BcPoA),SA(BcPDA)14

SA(TG)(BCPDA)isoSBMC

Schematic

Fig. 5Fig.5Fig. 6Fig.6Fig.9

1490 CLINICALCHEMISTRY,Vol.37. No.9, 1991

Table 2. TheoretIcal Amplifications wIth VariousBCPDA-Labeled Reagents

Amplification, -fold

15-1840

BSA. bovine serum albumin.

140-2101500-22504500-6750

Comparisons and Applications

My co-workers and I recently compared SA(BcPDA)14,SA(TG)(BCPDA)ise, and SA(TG)33(BcPDA) in variousnoncompetitive imniunoassay systems, using identicalconditions and protocols. Results are summarized inTable 3. The last-named reagent improved both signalamplification and detection limits by eight- to 26-foldcompared with SA(TG)(BCPDA)jso. Using the SBMC inan alpha-fetoprotein assay designed to maximize sensi-tivity, we were able to detect -300 000 molecules (-2amol) of alpha-fetoprotein (15), which is the lowestdetection limit reported for any time-resolved fluoro-metric assay (Table 4). The SBMC can be used in anyapplication in which a biotinylated reactant is used.

Documented applications other than immunoassay in-clude DNA hybridization2 and Western blotting. Cur-rently, the same reagent is being evaluated for North-ern blotting, DNA sequencing, and detection of PCRproducts. Other possible applications include flow-cy-tometry and inununohistochemistry.

Further Amplification Strategies

We have also extensively investigated strategies thatcould increase the biotinylation load. This increasedbiotinylation load, in combination with the SBMC,should yield even further total amplification of theproposed detection system. Among the many possibili-ties tested, two produced useful results.

The principle of the first approach is shown in Figure

9. Anti-lutropin detection antibodies are labeled withdigoxigenin, by use of an N-hydroxysuccinimide ester ofdigoxigenin. The immunological assay is performed asusual and the digoxigenin is further reacted with abiotinylated anti-digoxigemn monoclonal antibody. Wethen complete the assay by using the SBMC. A compar-ison of this assay with an assay involving a biotinylatedanti-lutropin detection antibody showed a 2.5-fold sig-nal amplification and improvement in detection limitwith the digoxigenin-labeled anti-lutropin antibody(Diainandis and Christopoulos, unpublished results).

The other approach is based on the use of horseradishperoxidase (EC 1.11.1.7)-catalyzed deposition of biotinson the solid phase as described by Bobrow et al. (16). Inthis approach, peroxidase is the label on the streptavi-din or the detection antibodies (Figure 10). After com-pleting the immunocomplex in the conventional manner,the enzyme substrate is added, H202, and the conjugatebiotin-tyramine. The peroxidase catalyzes the depositionof biotin-tyramine from solution onto the coating anti-body of the solid phase, thus markedly promoting theamount of biotin deposited in a manner linearly relatedto the amount of peroxidase present. These biotins arethen detected by the SBMC.3 With this system, we

2les TK, Diaznandis EP, Wilson G. Quantification ofnucleic acids on nitrocellulose membranes with time-resolvedfluorometry. Manuscript submitted.

3Christopoulos TK, Kitching R, Diamandis EP. Time-resolvedfluoroimmunoassays using enzymatic amplification. Manuscriptsubmitted.

Table 3. ComparIson of co

Analyt.

A-Labeled Streptavldln- Based Reagents In Various IDetection limit

mmunoassays

SA(BcPoa)14 SA(TGXBci’oA)iia, SAG)(acPoA)

Choriogonadotropin, mt.unit/L 1.00 0.32 0.03Lutropin,mt.unit/L 1.50 0.23 0.03Prolactin, ig/L 0.61 0.15 0.01Carcinoembryonic antigen, 1ug/L 0.36 0.11 0.01Thyrotropmn,milli-int.unitlL 0.45 0.09 0.01Follitropin,mt. unitlL 0.70 0.13 0.01Somatotropin,tg/L 0.27 0.08 0.003

Adapted from reference 14 by permission.

Table4. DetectIon Llmftsfor Alpha-Fetoproteln(AFP)AssaysIn VariousProtocols

CoatIngand-APP

antibody, pL

100100

Detection anti-APPantibody

Flat-bottom wellsMonoclonalPolyclonal

U-bottom wellsPolyclonalPolyclonalPolyclonal

Detection limIt

mol/L Meiecute&

ng/I. x 1O’ wail,x 10

12 20.2 60674.6 7.7 2309

Adapted from reference 15.

20 4.3 7.2 86310 5.5 9.2 5525 5.7 9.5 286

SBMC

because of an increase in background. In conclusion,,although some of the reagents described here are compli-cated, and considerable organic chemistry expertise isneeded to prepare and purify them, these reagents areextremely stable and easy to use. The reagents arepreferable to enzyme-based systems because they do notneed special conditions of temperature control, exacttiming, or extra steps for substrate incubation. In addi-tion, the system described here is completely contssmina-tion-free because the label (BCPDA) is a compound notpresent in nature-another advantage over both theEu3- and the enzyme-based assays. The system I havedescribed introduces amplification factors similar tothose achieved by the use of enzymes.

CLINICALCHEMISTRY,Vol.37, No.9, 1991 1491

Fig. 9. Haptenized detection antibodies with digoxigenin (DIG), andlinkage of the SBMC to the lmmunoconiplex by using biotinylatedanli-digoxigenin monoclonal antibodies

amplified the signal by 10- to 30-fold in comparisonwith biotinylated detection antibodies. However, the do-tection limit was improved by only about three- to sixfold

References1. Soini E, Hemmila I. Fluoroimmunoassay: present status andkey problems [Review]. Chit Chem 1979;25:353-61.2. HemmilA I. Fluoroimmunoassays and immunofluorometric as-says [Review]. Clin Chem 1985;31:359-70.

H,Ot

Biotin-Tyramins(B-I)

HAP

1_B 1_B i_B

YFIg. 10. Catalytic deposition of biotin-tyramine conjugate (B-i) onthe proteins-coated well by the action of horseradish peroxidase(HRP), in the presence of H202HAP is conjugated to either streptavidin or to detection antibody (this option isnot shown). The Immobilized B-I can be detected by use of the SSMC (notshown). Formore details see text and reference 16

3. Wieder I. Background rejection in fluorescence immunoasaay.In: Knapp W, Holubar K, Wick G, eds. Iznmunofluoreecence andrelated staining techniques. New York.:Elsevier, 1978:67-80.4. LeifRC, Clay SP, Gratzner HG, et al. Markers for instrumentalevaluation of cells of the female reproductive tract: existing andnew markers. In: Wied GL, Bahr GF, Bartels PH, ads. Theautomation of uterine cancer cytology. Tutorials of cytology. Chi-cago, 1975:313-44.5. Soini E, Hemmilfi I, Dahlen P. Time-resolved fluorescence inbiospecific assays. Ann Biol Clin 1990;48:567-71.6. Evangelista RA, Pollak A, Allore B, Templeton EF, Morton RC,Diamandis EP. A new europium chelate for protein labeling andtime-resolved fluorometric applications. Clin Biochem1988;21:173-8.7. Diamandis EP, Morton RC. Time-resolved fluorescence using aeuropium chelate of 4,7-bis(chlorosulfophenyl)-1,10-phenanthro-line-2,9-dicarboxylic acid (BCPDA). Labeling procedures and appli-cations in immunoasaaya. J Immunol Methods 1988;112:43-52.8. Gudgin-Templeton EF, Pollak A. Spectroecopiccharacteriza-tion of 1,10-phenanthroline-2,9-dicarboxylic acid and its com-plexes with europium(ffl): luminescent europium chelates usefulfor analytical applications in aqueous solution. J Lumin1989;43:195-205.9. Eeichstein B, Shami Y, Ramjeesingh M, Diamandis EP. Laser-excited time-resolved solid-phase fluoroimmunoassays with the neweuropium chelate 4,7-bis(chlorosulfophenyl)-1,10-phenanthroline-1,9-dicathoxylic acid as label. Anal Chem 1988;60:1069-74.10. Diamandis EP, Christopouloe TK. The biotin-(strept)avidinsystem. Principles and applications in biotechnology [Review].Clin Chem 1991;37:625-36.11. Diamandis EP, Morton RC, Reichstein E, Khoeravi MJ. Mul-tiple fluorescence labeling with europium chelators. Application totime-resolved fluoroimmunoassays. Anal Chem 1989;61:48-53.12. Rowley GL, Henriksson T, Louie A, et al. Sensitive fluoroim-munoassays for ferntin and 1gB [Tech Briefi. Cliii Chem1987;33:1563.13. Haralambidis J, Angus K, Pownall S, Duncan L, Chai M,Tregear W. The preparation of polyamide-oligonucleotide probescontaining multiple non-radioactive labels. Nucleic Acids Rae1990;18:501-5.14. Morton EC, Dianiandis EP. Streptavidin-based macromolecu-lar complex labeled with a europium chelator suitable for time-resolved fluorescence ixnmunoassay applications. Anal Chem1990;62:1841-5.15. Christopoulos TK, Lianidou ES, Diamandis EP. Ultrasensi-tive time-resolved fluorescence method for a-fetoprotein. ClinChem 1990;36:1497-502.16. Bobrow MN, Harris PD, Shaughnessy KJ, Litt GJ. Catalyzedreporter deposition, a novel method of signal amplification. Appli-cation to inimunoassays. J Immunol Methods 1989;125:279-85.