ANALYST, JULY 1992, VOL. 117 1061 Co-fluorescence Effect in Time-resolved Fluoroimmunoassays A Review Yong-Yuan Xu Analytical Chemistry Laboratory, Institute of Atomic Energ y, P.U. Box 275-88, Beijing 1024 13, China llkka A. Hemmila" Wallac Oy, P.O. Box 10, 20101 Turku, Finland Tim0 N.-E. Lovgren Department of Biochemistry, University of Turku, Turku, Finland Summary of Contents Introduction Time-resolved FI uoroi m mu noassays Co-fluorescence Effect 6-Di ketones Enhancing Ions Synergistic Agents Water-soluble Organic Solvents Buffers Preparation Enhancement Kinetics FI u o rescence Characteristics Time-resolved lmmunofluorimetry of FSH Double-label lmmunometric Assay of LH and FSH Quadruple-label lmmunoassay Composition of Co-fluorescence Enhancement Solution Co-f I uorescence En ha ncement Solutions Co-f luorescence En ha ncement in I m m u noassays Conclusions References Keywords: Time-resolved fluoroimmunoassa y; co-fluorescence enhancement; re view Introduction Time-resolved Fluoroimmunoassays In immunoassays the sensitivity is of prime importance, because often very low concentrations of analytes are present in the samples.In order to ensure the assay sensitivity, the antibody, antigen or its derivative has to be coupled with an easily detectable label. The conventional radioimmunoassays (RIA) and immunoradiometric assays (IRMA) utilize radio- active isotopes as the labels and are among the most sensitive and specific analytical techniques available today. The sensi- tivity of RIA, however, is still limited to a concentration range of 1 x 10-12-1 x 10-14 mol dm-3, and the relevant working range is often between 1 x 10-8 and 1 x 10-10 mol dm-3.The disadvantages of radioactive labels relate to their limited storage time and the legal problems and health hazards in their handling and disposal. The future use of radioisotopic techniques is also hindered by difficulties encountered in assay automation; automated RIA systems for routine diagnostic use have not emerged. Consequently, intensive research has been carried out with the aim of replacing the radioisotopes with stable, non-radioactive labels with a high specific activity. Fluorescent labels and fluorimetric detection represent one of the alternative non-isotopic techniques considered to have the potential to replace radioisotopic labels. The sensitivity of fluorescent labels, although theoretically very high, is, however, limited in routine immunoassay conditions by a high background interference. The background signal originates from various sources, such as fluorescent compounds in the * To whom correspondence should be addressed.sample, impurities in reagents and luminescent compounds in cuvettes and lenses, etc. For example, serum, the most widely used sample in immunoassays, causes very high background interference and, consequently, fluorimetric analysis in the presence of serum components (homogeneous assays) is limited to ymol dm-3-nmol dm-3 concentrations. The scatter- ing of the excitation light by sample constituents and solid-phase materials also gives rise to interference, especially when labels with a small Stokes' shift (less than 50 nm) are used.The high background limits the sensitivity of conven- tional fluoroimmunoassays by factors ranging from 100 to 1000.1 22 The use of delayed measurement of a long-lifetime fluores- cence makes it possible to separate the specific fluorescence of the label from most of the disturbing, unspecific background. In time-resolved fluorimetry the fluorescent label is excited by repeated short light pulses and the specific fluorescence is detected when a predetermined period of time (delay time) has elapsed after the excitation pulses. In order to be practical in routine immunoassay conditions, the fluorescent label employed should have as high a fluorescence as possible, produce emission preferably at a long wavelength and with a large Stokes' shift and, in particular, should have an excited- state lifetime clearly longer than the average duration of the background.1.3 The fluorescent properties of some tripositive lanthanide ions, and especially the chelates of Eu3+, Tb3+ and Sm3+, are particularly well suited for time-resolved fluorimetry. In these chelates the strong ion emission originates from an intra- chelate energy transfer, where the organic ligand absorbs the excitation radiation in the ultraviolet (UV) range and transfers1062 ANALYST, JULY 1992, VOL. 117 the excited energy through its triplet state to the emitting ion.4.5 The ligand field around the ion also prevents the quenching caused by coordinated water molecules which in aqueous solution tend to create an efficient deactivation route .6 The ion-specific emission appears at narrow banded lines at long wavelengths (Tb3+ 544 nm, Eu3+ 613 nm, Sm3+ 643 nm) with a long Stokes’ shift (230-300 nm).The most important feature in this context is the long fluorescence lifetime, ranging from 1 ps to over 2 ms, which makes it possible to apply time-resolved detection for the effective elimination of the background and to increase the sensitivity. The requirements set on a lanthanide chelate applied as a label in immunoassay relate to its stability, its fluorescence intensity in an aqueous environment, its hydrophilicity and the suitability for covalent coupling with antibodies. So far, the most successful method meeting these requirements is a fluorescence enhancement-based technique, marketed as DELFIA (Wallac, Turku, Finland).In this technique the lanthanide ion, generally Eu3+, is coupled to antibodies or antigens via bifunctional chelating agents to obtain stable but almost non-fluorescent labelled reagents.7.8 After completion of the heterogeneous immunoassay on a plastic surface (most often on microtitration strip wells), the label ion is dissociated by an acidic enhancement solution. The dissociated ion is simultaneously converted into a new, highly fluorescent chelate with P-diketones and a synergistic agent, trioctyl- phosphine oxide (TOPO), present in high excess in the acidic detergent solution.9~10 The enhancement solution forms opti- mum conditions for the chelate fluorescence, and a sensitivity of 1 x 10-14 mol dm-3 can be reached for Eu3+ detection in a sensitive time-resolved fluorimeter.3 A comprehensive list of the various DELFIA applications available for clinical immu- noassays, research and deoxyribonucleic acid (DNA) hybridi- zation, can be found in a recent review.” The ion-specific, narrow-banded emissions of the different lanthanide chelates, which, in addition, also show clearly different decay times, offer a tempting possibility to create double- or even multiple-label assays.10.11 The composition of the enhancement solution can be stipulated for the optimized measurement of either Eu3+ and Sm3+8710 or Tb3+,12 e.g., by changing the ligand from a fluorinated aromatic to a fluori- nated aliphatic P-diketone.By using this system (in this context called a direct fluorescence enhancement system, DFES), double-label based on Eu3+ and Sm3+*3914 or Eu3+ and Tb3+153 have been tested.In the direct system, the simultaneous use of three labels either results in a lower assay sensitivity or requires a two-step approach.17 Addition of the fourth label, Dy3+, to the direct system is hindered by its low emission quantum yield and, in particular, by its sub-micro- second decay time.” In the search for more sensitive methods enabling several analytes to be measured simultaneously with high sensitivity, the co-fluorescence type of enhancement (CFES) is potentially a very attractive approach. Co-fluorescence Effect The addition of yttrium or certain lanthanide ions can considerably enhance the fluorescence intensities of the P-diketone chelates of Eu3+ and Sm3+ when the chelates are present as a suspension in an aqueous solution.This type of fluorescence enhancement was first reported in 196418 and in 196719 in a study of Eu3+ (or Sm3+)-thenoyltrifluoroacetone (nA)-phenanthroline (Phen) and Eu3+ (or Sm3+)-7TA- collidine chelates coprecipitated in the presence of Gd3+ or Tb3+. Later it was shown that this type of fluorescence enhancement of Eu3+ or Sm3+-7TA chelates actually is an intrinsic fluorescence phenomenon named the ‘co-fluores- cence effect.’20-25 The fluorescence enhancement by CFES is based on an intermolecular energy transfer from the chelates of the emitting ion (such as Eu3+ and Sm3+). In all of the CFES systems studied both chelates have identical structures and are formed with f3-diketones (e.g., TTA) in addition to some synergistic ligands (e.g., TOPO or pyridine derivatives).The chelates formed have the structure Ln(L)3S1-2. The co-flu- orescence can be found in coprecipitates, in chelate suspen- sions and also in a micellar environment.22.24 In an actual solution, e.g., in benzene, there is no co-fluorescence en- hancement because the chelates exist as single molecules and the long distance between the chelates makes intermolecular energy transfer impossible. In an aqueous solution, however, van der Waals forces aggregate the hydrophobic chelates, which form tiny particles of controlled composition, and in the presence of a high excess of the enhancing chelates the emitting chelates are closely associated with a large number of enhancing chelates within the particles.The close contact allows efficient energy transfer from the latter to the former. The proposed energy transfer mechanism can be resonance energy transfer, collisional transfer of exciton migration. Ci and Lan24 pointed out that the energy transfer might be of exciton transfer type because it is independent of the acceptor concentration. The aggregates formed resemble micro- crystals, in which the structurally identical chelates are highly organized containing the emitting chelates as ‘impurities.’ The absorbed energy delocalizes through the chelate matrix, formed by the enhancing chelates, and ends up in the emitting chelates which produce the emission.22.24 The mechanism of the energy transfer process in CFES is illustrated in Fig.1. First, the organic ligand of both chelates (such as Ln-=A3) absorbs the excitation light and its electrons are raised from the singlet ground state (So) to one of the vibrational multiples of the excited state (S1). Subse- quently, the energy falls rapidly to the lowest level of S1 through non-radiative deactivation processes, and by inter- system crossing to the central ion-stabilized triplet state (T1). In the luminescent chelate, the excited energy at T1 is transferred by intrachelate energy transfer to the lowest resonance level of the emitting ion (e.g., Eu3+), whereafter the ion undergoes a radiative transition resulting in a characteristic line-type emission characteristic of the ion. In the aggregates most of the excitation light is absorbed by the chelates of non-emitting ions, present in high excess.The W I 2 =I s. intersvstem -+1 crossing # I-- I ntra molecu la r 1- Y -0 Inter m o lecu I a r energy transfer a t a v) a -3 7F U’ - Ligand Eu3+ Gd3+ Ligand - Fluorescent chelate Enhancing chelate . V Fie. 1 Schematic diagram of the energy flow in ligand-sensitized enhancing ion (such as Gd3+ and Tb3+) to the chelates of the direct and indirect (co-horescence) c h e k e luminesckeANALYST, JULY 1992, VOL. 117 1063 excited energy at the ligand triplet state cannot be transferred to its central ion, e.g., with Gd3+, because its resonance energy level is located at a much higher level. Because no radiative excited-state deactivation processes exist within the enhancing chelates, the stabilized triplet state transfers the energy by intermolecular energy transfer of the nearby chelates in the aggregated particles, thus producing a con- siderably enhanced luminescence intensity.Composition of Co-fluorescence Enhancement Solution The hitherto studied co-fluorescence-sensitized enhancement systems, similar to most of the direct systems, are based on P-diketones as the energy-absorbing ligands forming tris-(3- diketonato chelates with the trivalent lanthanide ions. In addition to the P-diketones, synergistic ligands, most fre- quently N-heterocyclic Lewis bases, are utilized as neutraliz- ing and water molecule-replacing ligands. Analogously to the direct systems, detergents have been found to stabilize the system and sometimes organic solvents have a positive effect on the fluorescence.P-Diketones P-Diketones have gained the widest use as the ligands to increase lanthanide fluorescence. As bidentate chelating agents, they form relatively stable chelates. The six-mem- bered ring involved in the chelate structure directly absorbs the excitation light and efficiently transfers the energy to the chelated ion. Aromatic P-diketones containing a triflu- oromethyl group are commonly used for the fluorimetric determination of Eu3+ and Sm3+, e.g., after extraction into organic solvents.26~27 These ligands are also used in most of the DELFIA-type time-resolved fluoroimmunoassays.8-11 Fluori- nated aliphatic P-diketones have the excited triplet state at a higher energy level and, therefore, can also be used for sensitizing Tb3+ and Dy3+ luminescence,11.12 and can be applied to the simultaneous detection of Eu3+, Tb3+, Sm3+ and Dy3+.Thenoyltrifluoroacetone is one of the most commonly used P-diketones to detect Eu3+ and Sm3+, used either in organic solvents or in aqueous solution .26,27 Thenoyltrifluoroacetone also promotes a strong co-fluorescence effect, and so far most of the co-fluorescence experiments have been performed using TTA as the fluorogenic ligand.19-25 We have found, however, that other P-diketones also show the co-fluorescence effect under appropriate conditions28.29 (Table 1), and by applying, e . g . , aliphatic p-diketones the co-fluorescence enhancement system can be expanded for the determination of Tb3+ and Dy3+.29330 The co-fluorescence enhancement efficiency greatly depends on the structure of the P-diketone.According to our results, the P-diketones listed in Table 1 can be classified into three major categories: (1) (3-diketones exhibiting a strong co-fluorescence effect, such as TTA, BTA and PTA; (2) P-diketones showing a less intense co-fluorescence effect, e.g., FTA, FBTA, TFMH, DPM, HFAcA, PFH, PFDMH and HFDMO; and (3) P-diketones which show only a weak interchelate energy transfer under the conditions tested, such as DBM, PFPP, DFBM, (3-NTA, AcA, TFAcA, TFTD and PFTD . Although a quantitative correlation between the structure of the 6-diketone and the strength of the co-fluorescence effect has not been verified, some qualitative consistencies have been observed. Generally, a CF3 group is needed for the enhancement of both co-fluorescence and direct fluorescence, but a higher level of fluorination did not increase the co-fluorescence efficiency.A long fluorocarbon side-chain (R1, in TFTD and PFTD) or a bulky aromatic structure (DBM, DFBM, P-NTA), however, had a negative effect on the interchelate energy transfer. Accordingly, the optimum ligand in the direct enhancement system, P-NTA,s showed only a weak co-fluorescence effect. The excitation of Eu3+ and Sm3+ is accomplished more efficiently through aromatic P-diketones, and accordingly the highest sensitivities are obtained with P-diketones such as TTA or BTA. On the other hand, multi-label assays, in which more than two analytes are to be measured simultaneously, require an enhancement system composed of aliphatic p-diketones.160 I 1 40 2 20 L L 0 20 40 60 80 100 120 140 160 PTNymol dm-3 Fig. 2 Effect of PTA on the fluorescence of 40 pmol dm-3 E d + (W) and 25 pmol dm-3 Tb3+ (A) in the PTA-Phen-Y3+-Triton-ethanol system Table 1 Structures of j3-dikctones (general formula R*COCH2COR2) with co-fluorescence enhancement effect Aromatic j3-diketones- R' Thenoyltrifluoroacetone (TTA) Benzoyltrifluoroacetone (BTA) 2-Furoyltrifluoroacetone (FTA) p-Fluorobenzoyltrifluoroacetone (FBTA) P-Naphthoyltrifluoroacetone (P-NTA) 1,l ,1,2,2-Pentafluor0-5-phenylpentane-3,5-dione (PFPP) Dibenzoylmethane (DBM) Di-p-fluorobenzoylmethane (DFBM) Pivaloyltrifluoroacetone (PTA) l,l,l-Trifluoro-6-methylheptane-2,4-dione (TFMH) Dipivaloylmethane (DPM) 1,1,1,5.5,5-Hexafluoroacetylacetone (HFAcA) 1 ,l ,1,2,2-Pentafluorohexane-3,5-dione (PFH) 1,1,1,2,2-Pentafluoro-6,6-dimethylheptane-3,5-dione (PFDMH) 1.1,1.2,2,3,3-Heptafluoro-7,7-dimethyloctane-4,6-dione (HFDMO) 1,1,1,2,2-Pentafluorotetradecane-2,4-dione (PFTD) 1,1,1-Trifluorotridecane-2,4-dione (TFTD) l,l,l-Trifluoroacetylacetone (TFAcA) Acetylacetone (AcA) Aliphatic j3-diketone.s-1064 ANALYST, JULY 1992, VOL.117 The effect of the P-diketone concentration on the co-fluor- escence enhancement is exemplified in Fig. 2, where the effect of PTA on both Eu3+ and Tb3+ emissions in the CFES system is shown. The optimum concentration range can either be wide (as for PTA in Eu3+ detection) or narrow (as for PTA in Tb3+ detection). The optimum concentration of 6-diketones is almost always sufficiently high to ensure ternary tris-chelate formation with the enhancing ion (concentration of P-dike- tone more than three times that of the enhancing ion).Enhancing Ions The enhancing ions generally applied are Gd3+, Tb3+, Lu3+, La3+ and Y3+. Under some conditions, a weak co-fluores- cence enhancement effect is also obtained with Yb3+ and Dy3+. In the co-fluorescence enhancement system the enhanc- ing ion used must not have excited 4f or 4d levels situated below the excited triplet level of the P-diketone used. Hence the energy absorbed by these chelates cannot be dissipated through these non-existing energy levels. Instead, the energy is transferred to the fluorescent ions through an intermol- ecular energy transfer. Enhancing ions are needed to provide the high molar excess of the triplet sensitizer to ensure a linear response for the acceptor detection.The high concentration is also needed to create the aggregate particles inside which the efficient energy transfer is possible. In the mA-based co-fluorescence system the fluorescence intensity of Eu3+ increased with increasing concentration of Y3+ (Fig. 3), and the maximum fluorescence intensity was obtained at a Y3+ concentration of about 5.0 pmol dm-3. The optimum concentration was not dependent on the Eu3+ concentration. A similar result was also obtained with Sm3+ (data not shown). Hence no constant molar ratio of Eu3+ (or Sm3+) to Y3+ exists in the particulate chelate structures, suggesting that there are no defined mixed chelate structures in the particles. Taking into account the bidentate nature of 107 r I r 1,106 v) 4- 105 ; 104 E 103 0 \ a 0 ii 102 I I I 0 5 10 15 20 Y3+/1.~mol dm-3 Fig.3 Effect of Y3+ on the fluorescence of 0.5 pmol dm-3 (V), 5 pmol dm-3 (A) and 50 pmol dm-3 (@) Eu3+ in the TTA-Phen-Y3+- Triton system Table 2 Enhancement factors* of Eu3+ and Sm3+ fluorescence by different enhancing ions in the Eu3+ (or Sm3+)-TTA-Phen-Ei-Triton X-100 system Fluorescence enhancement factor Enhancing ion Gd3+ Y3+ Tb3+ Lu3+ La3+ Yb3+ Dy3+ Eu3+ Sm3+ 740 766 526 81 1 493 757 504 729 135 168 45 20 15 19 * The enhancement factor is defined as the ratio of the fluorescence intensity in the presence of an enhancing ion to that in its absence. P-diketones, the formation of mixed ligand chelates would be most unlikely. Actually, the chelate stoichiometry of 1 : 3 : 1-2 (Ln : P-diketone : synergistic ligand) has been verified for co-fluorescent aggregates.22.24 The co-fluorescence effect is caused only by the intermolecular energy transfer between the two chelate types.The enhancement factors obtained with the different enhancing ions (Ei) in the Eu3+ (or Sm3+)-?TA-Phen-Ei- Triton X-100 systems are given in Table 2. The lanthanides Gd3+, Tb3+ and Lu3+ and the non-lanthanide Y3+ gave the highest enhancement factors in the TTA-based CFES studied. An additional aspect in the choice of the enhancing ion is the inherent contamination-even trace amounts of the emitting lanthanides as impurities cause a very high background. In the respect the non-lanthanide Y3+ is to be preferred. Synergistic Agents The strong fluorescence intensity of the lanthanide 6-diketone chelates requires a non-aqueous chelate environment.The three P-diketone molecules in the tris-chelate occupy only six of the nine available coordination sites of the ion, which still remains sensitive to quenching by water. Halverson et a1.31,32 introduced the concept of an ‘insulating sheet,’ in which a synergistic ligand (or a synergistic agent) was involved in the chelate structure to replace water molecules from the ligand field. It also acts as a shield, protecting the chelate from external interactions and thus efficiently reducing non-radia- tive energy degradation. The chelate formed has the structure Ln3+ (P-diket~ne)~(L)~-~. According to our investigation, the synergistic ligands applicable for co-fluorescence enhancement are 1,lO-phenan- throline (Phen) and its derivatives, e.g., 4,7-(or 5,6)- dimethyl-1 ,lo-phenanthroline, 4,7-diphenyI-l,lO-phenan- throline and 2,9-dimethyl-4,7-diphenyl- 1 , 10-phenanthroline, pyridine derivatives such as 2,2‘-dipyridyl (DP) , 2,2’-dipyr- idylamine, 2,4,6-trimethylpyridine, 2,2’ : 6’,2”-tetrapyridine and 1,3-diphenylguanidine.Trioctylphosphine oxide is the most effective synergistic agent in the direct enhancement solutions and has been applied in some CFES studies.21 In our systems, however, it almost totally hindered the interchelate energy transfer within the CFES aggregates. Usually, the fluorescence on the lanthanide chelate increases with increasing concentration of the synergistic ligand until a maximum and almost stable fluorescence is reached.The effect of the Phen concentration on the fluorescence of Eu3+ and Sm3+ chelates in BTA-Ys+-based CFES is shown in Fig. 4.28 Detergents have a positive effect both on the direct system and on the co-fluorescence-enhanced system. The micellar environment protects the chelates against non-radiative processes, provides a non-aqueous environment for the chelate and ensures an environment in which organized 2000 v) v) 4- 5 1500 s m I 0 a C a y 1000 500 L 0 3 U - 0 20 40 60 80 100 Phen/prnol dm-3 Fig. 4 Effect of Phen on the fluorescence of50 pmol dm-3 Eu3+ (V) and 3.5 nmol dm-3 Sm3+ (A) in BTA-Phen-Y3+-Triton X-100 based CFESANALYST, JULY 1992, VOL. 117 1065 crystalline stuctures can be formed. Detergents can also solubilize the particles and stabilize the solutions by prevent- ing the sedimentation of the particles.The type of detergent used in CFES varies according to the system studied and positively or negatively charged and non-ionic detergents have been tested. Ci and Lan24 reported that an increasing concentration of detergent, e.g., Triton X-100, can even Preparation a quenching effect. A typical effect of pH on the co- fluorescence of Eu3+ and Sm3+ in BTA-Phen-Y3+-Triton- based CFES is shown in Fig. 7.28 Co-fluorescence Enhancement Solutions solubilize the aggregate particles. However, the effective concentration of detergents is well below the critical micelliza- tion concentration (c.m.c.). We have studied the effect of non-ionic detergents, Triton X-100, Triton N-101 and Triton X-450, on the co-fluorescence system.The types of effect caused by the detergent on the direct fluorescence and the co-fluorescence enhancement are different. An increasing concentration of the detergent did not change the fluorescence of the Tb3+-PTA chelate after reaching the optimum level in DFES consisting of lW+-PTA-TOPO-Triton X-10O.l2 In the respective co-fluorescence system based on Tb3+-PTA- Phen-Y3+-Triton X-100, the intensity gradually decreased beyond the optimum Triton X-100 concentration of about 0.06% .29 Obviously, at the higher Triton concentrations the chelate densities within the micelles are decreased, and eventually the fluorescent chelates and enhancing chelates may even be separated into different micelles if the micelle concentration exceeds that of the chelates. Under such conditions interchelate energy transfer is impossible.A small concentration of detergent also stabilizes the co-fluorescence system. The tiny particles in the detergent buffer remain in suspension for a longer time. The effect of different Triton X-100 concentrations on the stability of the Eu3+-BTA- Phen-Y3+-Triton X-100 system is shown in Fig. 5. Water-soluble Organic Solvents Some water-soluble organic solvents, such as ethanol, pro- panol, dimethyl sulfoxide, 2-methoxyethanol and ethane-l,2- diol, can improve the co-fluorescence intensity. In the Tb3+-PTA-Phen-Y3+-Triton X-100-ethanol system, the flu- orescence of the Tb3+ chelate is enhanced 186-fold with the addition of ethanol (Fig. 6).29 It is likely that the presence of ethanol favours the formation of optimally sized particles.Buffers Unlike the direct enhancement system, which can be opti- mized both to dissociate the ion by the acidic pH and to create the optimally emitting chelates simultaneously,8 the optimum and stable enhancement in the co-fluorescence system requires careful adjustment of pH near the neutral range. Acetate, hexamine and tris(hydroxymethy1)methylamine (Tris) can be used, whereas phosphate and citrate buffers have 500 1 1 I I I 0 20 40 60 80 Ti me/m i n The final CFES is stable for only a limited period of time because the solution contains tiny particles present in the micellar environment. The optimum pH of the CFES is near neutral. Consequently, CFES cannot be applied directly in the dissociation-enhanced lanthanide fluoroimmunoassay prin- ciple.Instead, the CFES components are prepared, stored and used in two separate solutions. The first solution (E,) contains the (3-diketone, the enhancing ion, the detergent and, sometimes, a water-soluble organic solvent, solubilized in an acidic buffer (pH 3.1-3.5). The second solution (Eb) contains the synergistic ligand and a basic buffer. When CFES is employed, e.g., in immunoassays based on antibodies labelled with lanthanide chelates, the solution E, is first used to dissociate the ions and thereafter the solution E b is added (usually one tenth of the E, volume) to raise the pH to the optimum level and to allow the formation of the particulate structures necessary for the co-fluorescence enhancement.The compositions of five different CFESs are given in Table 3. The sensitivities obtained with the CFES depend not only on the signal levels obtained but also on the signal-to-noise ratio. Consequently, the fluorescence intensity should be high, the decay time long and the background level low and reproducible. The background obtained is derived partly from the instrumental background and the photoluminescence of the solid-phase material (polystyrene) and to a great extent from the purity of reagents (contamination level). In order to keep the background at a low level, the follovving precautions VJ t U 3 40 8 m O 30 0 10 20 30 40 Ethanol (%) Fig. 6 Effect of ethanol on the fluorescence of 25 pmol dm-3 Tb3+ in the PTA-Phen-Y3+-Triton X-1Okthanol system at different Triton X-100 concentrations: 0 (+), 0.006 (V), 0.03 (A), 0.06 (M) and 0.12% (0) 2000 1 v) VJ - 5 1500 0 m 0 Q, c Q, u y 1000 500 0 L L - 0 4 5 6 7 8 PH Fig.5 Stability of fluorescence, on standing, of 50 pmol dm-3 Eu3+ in BTA-Phen-Y3+-Triton X-100 based CFES with different Triton X-100 concentrations: 0 (0), 0.005 (A), 0.01 (V), 0.02 (0) and 0.04% (M) CFES system Fig. 7 Effect of pH on the fluorescence of 50 pmol dm-3 E d + (A) and 3.5 nmol dm-3 Sm3+ (V) in the BTA-Phen-Y3+-Triton based1066 ANALYST, JULY 1992, VOL. 117 Table 3 Co-fluorescence enhancement characteristics of different systems CFES Shaking time/min Composition Fluorescent ion Ea* Eb TPY TPYE BPY PPY E PDY Eu3+, Sm3+ 70 pmol dm-3 TTA 7.5 pmol dm-3 Y3+ 0.98 cm3 dm-3 HOAc 30 pmol dm-3 TTA 1.5 pmol dm-3 Y3+ 0.06% TX-100 20% ethanol 0.96 cm3 dm-3 HOAc 60 pmol dm-3 BTA 8.5 pmol dm-3 Y3+ 0.98 cm dm-3 HOAc 70 pmol dm-3 PTA 7.5 pmol dm-3 Y3+ 0.06% TX-100 30% ethanol 1.04 cm3 dm-3 HOAc 100 pmol dm-3 PTA 3 pmol dm-3 Y3+ 2 cm3 dm-3 HOAc 0.075% TX-100 Eu3+, Sm3+ Eu3+, Sm3+ 0.02% TX-100 Eu3+, Sm3+ Tb3+, Dy3+ Eu3+, Sm3+ Tb3+, Dy3+ 0.0006% TX-100 * TX-100 = Triton X-100; HOAc = acetic acid.1.75 mmol dm-3 Phen 8 0.2 mol dm-3 Tris 0.4 mmol dm-3 Phen 7 0.2 mol dm-3 Tris 0.5 mmol dm-3 Phen 0.2 mol dm-3 Tris 0.5 mmol dm-3 Phen 0.2 mol dm-3Tris 5 mmol dm-3 DP 0.375 mol dm-3 Tris 80% ethanol 1 7 15 Waiting time/min 10 10 0 10 20 need to be taken. All reagents used to prepare the CFES must be very pure. Non-fluorescent lanthanides often contain high levels of fluorescent lanthanide ions, causing a high back- ground.A trace contamination level of 0.0001% in yttrium oxide (99.9999% pure) was found to be an acceptable level. All laboratory ware used in the preparation of CFES must be made of plastic and must be thoroughly washed with a suitable washing solution before use. Frequent rinses have to be carried out, e.g., with E, solution. During use and storage, special care must be taken in order to avoid contamination of the solution. Enhancement Kinetics In the co-fluorescence system, the ternary chelate structures exist as tiny particles which tend to precipitate out from the solution after prolonged standing. Standing of ready CFES for 24 h resulted in sedimentation of the particles and disappear- ance of the fluorescence from solution. The precipitated particulate material exhibits very strong fluorescence, whereas most of the fluorescence in solution disappears.Evidently both the fluorescent chelates and the enhancing chelates are in the same particles. The total enhancement time depends on the kinetic stability of the chelate used for labelling and on the time required for aggregate formation. When N1-(p-isothiocyanatobeny1)- diethyIenetriamine-Nl,W,N3,W-tetraacetic acid is used as the chelate for labelling of the reagents with the fluorogenic lanthanide,’ the ion dissociates within 3-5 min in the E, solution. After the addition of Eb the development of the final fluorescence depends on the speed of particle formation, which varies for different co-fluorescence enhancement systems (Table 3). Fluorescence Characteristics The excitation and emission spectra of Eu3+, Tb3+, Sm3+ and Dy3+ chelates in PPYE solution are shown in Fig.8. The excitation maximum of the PTA chelate is red shifted from the 295 nm obtained in DFESl2 and shows two maxima at 300 and 315 nm. A similar red shift has been reported with Eu-TTA chelates in CFES.24 The strongest emission peaks are found at 612, 544, 647 and 574 nm for Eu3+, Tb3+, Sm3+ and Dy3+ chelates, respectively, and are assigned to the transitions of Table 4 Fluorescence properties of lanthanides in CFES Excitation Emission Decay Composition Ion (max. )/nm (max .)/nm time/ys TPY Eu3+ Sm3+ TPY E Eu3+ Sm3+ BPY Eu3+ Sm3+ PPY E Eu3+ T b 3 + Sm3+ Dy3 + PDY Eu3+ Tb3+ Sm3+ Dy3 + 365 358 365 364 333 337 315 312 315 316 312 312 312 313 612 648 612 a 9 612 647 612 544 647 574 612 545 647 574 1062 96 730 96 764 79 820 323 88 27 948 239 48 11 5D0-+7F2 of Eu3+, 5D4-+7F5 of Tb3+, 4G5/2-+6H9/2 of Sm3+ and 4F9/2-+6H13/2 of Dy3+.The fluorescence properties of the chelates in the five CFES are given in Table 4. The excitation wavelength mainly depends on the b-diketones used and the emission wavelength completely on the fluorescent ion. The different ions also exhibit their typical decay times which, to some extent, depend on the ligand field around the ion. The measurement conditions, fluorescence intensities, back- grounds and detection limits of the four fluorescent lantha- nides measured in different CFESs are listed in Table 5. The calibration graphs exhibit wide linear responses with respect to emitting ion concentrations from 1 x 10-13 to 1 x 10-7 rnol dm-3.The detection limits with TPYE CFES were 2.5 X 10-15 mol dm-3 for Eu3+ and 8.9 x 10-14 rnol dm-3 for Sm3+, which are 15 and 37 times lower than those obtained using the respective DFES (3.9 x 10-14 rnol dm-3 for Eu3+ and 3.3 X 10-12 rnol dm-3 for Sm3+).* Regardless of both the spectral and temporal resolutions of the emissions of different lanthanides, minor spectral over- lapping occurs between some of the emissions. This needs to be corrected in simultaneous determinations in which the whole dynamic range of each ion has to be utilized. The signal cross-talk figures between the four fluorescent lanthanides in PPYE-based CFES are given in Table 6. The relatively broad emission peak of Sm3+ (Fig.8) causes a high signal interfer- ence in the Dy3+ channel.ANALYST, JULY 1992, VOL. 117 1067 Table 5 Measurement parameters and fluorescence of the lanthanides in CFES Measurement parameters Composition Ion TPY Eu3+ Sm3+ TPYE E u ~ + Sm3+ BPY E u ~ + Sm3+ PBYE E u ~ + n 3 + Sm3+ Dy3+ PDY E u ~ + Tb3+ Sm3 + Dy3+ Cycling/ ms 2 1 2 1 2 1 2 1 1 1 2 1 1 1 Delay/ ms 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.4 0.05 0.05 0.5 0.15 0.05 0.02 Counting/ ms 1.5 0.15 1.5 0.2 1.5 0.15 1.5 0.5 0.2 0.1 1.5 0.4 0.2 0.1 Excitation/ nm 310-400 310-400 310-400 310-400 310-400 310-400 250-400 250-400 250-400 250-400 250-400 250-400 250-400 250-400 Emission/ nm 613 643 613 643 613 643 613 545 643 573 613 545 643 573 Fluorescence for 1 nmol dm-3 Ln3+/l@ counts Background/ S-1 counts s-1 84 130 5 700 602 180 83 900 1456 608 85 41 940 1 860 324 400 3 640 760 1 880 4 700 12 410 20 6 200 6 846 lo00 2 983 2 400 17 100 25 5 800 Detection limit/ pmol dm-3 0.015 0.12 0.0025 0.089 0.0043 0.11 0.035 0.34 7.9 46.0 0.019 0.27 3.8 20.0 Table 6 Spectral interference* in the co-fluorescence measurement of lanthanides in CFES containing PPYE Ion to be measured Interfering ion E u ~ + m3+ Sm3+ Dy3+ Eu3+ - 0.17 0.32 0.33 Tb3+ 3.36 - 0.49 1.69 Sm3 + 1.76 0.00 - 5.39 Dy3+ 0.00 0.00 0.31 - * Interference percentage is defined as the ratio of the signal of the interfering ion obtained in the measuring channel to that of the ion to be measured. Co-fluorescence Enhancement in Immunoassays Time-resolved Immunofluorimetry of FSH A solid-phase-based non-competititive time-resolved flu- oroimmunoassay of follicle-stimulating hormone (FSH) was used as a model to assess the sensitivity obtainable with a CFES.Microtitration strip wells were coated with P-FSH specific monoclonal antibodies and a-specific antibody was labelled with Eu3+ [17 Eu3+ per molecule of immunoglobulin G (IgG)]. In the first step, samples and standards were incubated in the wells for 1 h at room temperature in the DELFIA assay buffer. In the second step, after washing, the wells were incubated with 5 ng of Eu3+-labelled anti-a-FSH (1 h), then washed six times and the bound fraction of Eu label was dissociated by shaking the strips for 3 min with 200 mm3 of TPY solution E,. The fluorescence of the label was enhanced by solution Eb (20 mm3 per well) during 8 min of shaking.The solution was allowed to stand for an additional 10 min, then the fluorescence was measured with a time-resolved spectro- fluorimeter. A direct system (commercial DELFIA enhance- ment solution) was used as a reference. The dose-response curves of FSH using the two different enhancement solutions are shown in Fig. 9. In the assay CFES gave clearly higher signal levels but the sensitivity of the assay was comparable to that of DFES. In this instance the sensitivity was more dependent on the non-specific binding property of the labelled antibody than on the detection sensitivity of the label. Double-label Immunometric Assay of LH and FSH Depending on the P-diketones used, the CFESs can be applied in double-label time-resolved fluoroimmunoassays using either Eu3+ and Sm3+ or Eu3+ and Tb3+ as the labels.The simultaneous assay of luteinizing hormone (LH) and FSH has been chosen as a model for double-label a~say.15~28~29 The assays were based on (3-specific anti-FSH and anti-LH 4- '! .- [ii / \ > 4- 'z (a) a, C Q, C a, tn 1c a, .- L! 260 280 300 320 340 360 4- I I I a, a > tn a, C a, 4-4 .- 4- .- h,,/nm I e m h m Fig. 8 Tb'+ (c), Sm3+ (d) and Dy3+ (e) chelates in PPYE-based CFES Excitation spectrum (a); and emission spectra of Eu3+ (b),ANALYST, JULY 1992, VOL. 117 L 1068 107 c I v) v) w = 106 s 8 2 v) 105 . W v) 3 - Y- w LL u 104 105 Y I v) v) w 104 =I 8 . 0 C W v) 2 103 3; n 3 - W LL 102 0.1 1 10 FSH/U dm-3 Fig. 9 Dose-response curves of time-resolved immunofluorimetry of FSH measured with use of TPY-based CFES (0) and P-NTA-TOPO- Triton-based DFES (m) 0.1 1 10 100 LH or FSH/U dm-3 Fig.10 Dose-response curves of double-label time-resolved immu- nofluorimetric assay of LH with Eu3+ (round symbols) and FSH with Sm3+ (square symbols), when measured with TPY-based (solid symbols) or BTA-based CFES (open symbols) Table 7 Comparison of double-label and single-label immunoassays of LH and FSH Sample SensitivityIU dm-3 volume/ Assay Labels mm3 LH FSH RIA* TO, 1251 200 24.4 1.5-3.6 TR-IFMAT Eu3+, Tb3+ 100 0.1 1.0 TR-IFMA with TR-IFMA with CFES (TPY) Eu3+, Sm3+ 25 0.045 0.029 CFES (PPYE) Eu3+, Tb3+ 25 0.024 0.017 DELFIA (DFES)$ Eu3+ - 0.02 - DELFIA (DEFS)$ Eu3+ - - 0.01 * Ref. 34. 7 Ref. 15. 3 Ref. 35. antibodies labelled with either Sm3+ or Tb3+. The microtitra- tion plate wells were coated with a common antibody against the a-subunit of both hormones.In both of the model assays the coated wells were incubated with 25 mm3 of samples or standards in 200 mm3 of assay buffer for 1 h with constant shaking. After three washings the wells were further incu- bated either with a mixture of 25 ng of E$+-anti-P-LH and 500 ng of Sm3+-anti-P-FSH, or a mixture of 25 ng of Ed+-anti-P-LH and 100 ng of Tb3+-anti-P-FSH monoclonal antibodies (1 h at room temperature with constant shaking). The label fractions immunologically bound on the solid c I v) ~1000 0 7 100 w Z t Y- 0 1 E l o t l 5 h 10 E n 5 % 2 1 10 100 - LL 0.1 100 TSH/pU cm-3 I(b) 40 2o t 0 I 10 100 . - I I v) C 3 $100 8 0 10 2 O 1 8 m F v) *- Q) v) 8 0.1 = 10 c F 2,1000 s fn + a m 100 I 0 r .0” Y- o 10 W C 8 $ 1 W - L 17-cu-OHP/nmol dm-3 100 1000 lRT/pg dm-3 l5 10 E n 5 g 0 1 25 20 15 E f3 10 g I 5 I I I I0 10 100 1000 CK-MM/U dm-3 Fig. 11 Dose-response curves and precision profiles of quadruple- label time-resolved immunofluorimetric assay of (a) TSH with Eu3+ ; (b) 17wOHP with Tb3+; (c) IRT with Sm3+; and ( d ) CK-MM with Dy3+ surfaces were dissociated with 200 mm3 of E, and the co-fluorescence was enhanced with 20 mm3 of Eb solution. The dose-response curves of the assays obtained with TPY- or BPY-based CFES (Eu3+ and Sm3+ as the labels) are shown in Fig. 10.28 A comparison of four different double-label immunoassays of LH and FSH, one radioimmunoassay , one time-resolved fluoroimmunoassay with PTA-TOPO-based DFES, two assays with CFES and the respective single-label assays is given in Table 7.The sensitivities with CFES were considerably higher (5-50-fold) than those obtained with previous DFES-based double-label assays. Quadruple-label Immunoassay The applicability of CFES containing an aliphatic P-diketone (PPYE) has also been tested in a quadruple label assay of four analytes used for neonatal screening: thyroid stimulatingANALYST, JULY 1992, VOL. 117 1069 hormone (TSH), immunoreactive trypsin (IRT), creatine kinase MM (CK-MM) and 17-a-hydroxyprogesterone (17a- OHP) .33 Three of the assays were based on non-competitive sandwich-type assays with monoclonal antibodies and the assay of 17a-OHP on a competitive assay with secondary antibodies on the solid surface.The antibodies, anti-TSH, anti-IRT and anti-CK-MM, were labelled with Eu3+, Sm3+ and Dy3+, respectively. The antigen, 17a-OHP, was directly labelled with Tb3+. The assay was performed as a one-step assay in microtitra- tion strip wells coated with a mixture of four antibodies. The samples were blood spots dried on a filter disc (diameter 5 mm). Standardization was performed with a dried blood spot standard for TSH and liquid standards for the others. The amounts of tracers used were 50 ng of Eu3+-anti$-TSH, 100 ng of Sm3+-anti-IRT, 250 ng of Dy”-anti-CK-MM and 0.09 pmol of Tb3+-17wOHP. The primary rabbit-anti-17-a-OHP antibody was used at a dilution of 1 : 7500. After overnight incubation at 4 “C and six washings, the bound fractions of the labels were dissociated by shaking for 4 min with 200 mm3 of solution E, (PPYE) and the co-fluorescence was enhanced by adding 20 mm3 of solution Eb for 10 min.After waiting for 10 min the fluorescence was measured on the time-resolved fluorimeter using appropriate parameters (Table 5 ) . The dose-response curves and precision profiles of TSH, 17a- OHP, IRT and CK-MM are shown in Fig. 11. The sensitivity of the assay was 0.1 pU cm-3 for TSH, 3 nmol dm-3 for OHP, 2 pg dm-3 for IRT and 4 U dm-3 for CK-MM. Conclusions Time-resolved fluoroimmunoassays using lanthanides as labels have already gained wide acceptance in a variety of routine clinical and research applications. When using the direct fluorescence enhancement solution (DELFIA) based on p-NTA and TOPO, a sensitivity of 8 x 10-18 mol of Eu3+ per well has been obtained.Although the clinical range of most analytes does not require the utmost sensitivity, the improved sensitivity is needed for certain analytes. There is also a specific need to improve the sensitivities in multi-label assays, where a simple DFES can be applied only for two sensitive assays simultaneously. The co-fluorescence enhance- ment effect has offered a way to increase further the sensitivity of TR-FIA. The improvement is based on the additional energy transfer creating an amplified excitation efficiency of the (3-diketone chelates. In CFES a sensitivity as high as 5 x 10-19 mol of Eu3+ per well is achieved. The further delayed emission as a result of the additional energy transfer processes taking place prior to the emission also makes the fluorimetric measurement of Dy3+ possible in the present time-resolved fluorimetry .Hence CFES permits simple and sensitive single-, double-, triple- and quadruple-label immunoassays, a feature unique in comparison with other non-istotopic label tech- niques. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Wieder, I., in Immunofluorescence and Related Staining Tech- niques, eds. Knapp, H., Holubar, H., and Wick, G., Elsevier North-Holland Biomedical Press, New York, 1978, pp, 67-80. Soini, E., and Kojola, H., Clin. Chem. (Winston-Salem, N.C.), 1983, 29, 65. Weissman, S. I., J. Chem. Phys., 1942, 10, 214. Crosby, G., Whan, R., and Alire, R., J . Chem. Phys., 1961,34, 743. Bryden, C., and Reilley, C., Anal.Chem., 1982, 54, 610. Mukkala, V.-M., Mikola, H., and Hemmila, I . , Anal. Bio- chem., 1989, 176, 319. Hemmila, I., Dakubu, S . , Mukkala, V.-M., Siitari, H., and Lovgren, T., Anal. Biochem., 1984, 137, 335. Lovgren. T., Hemmila, I., Pettersson, K., and Halonen, P., in Alternative Immunoassays, ed. Collins, W., Wiley, Chichester, Hemmila. I., Scand. J. Clin. Lab. Invest., 1988, 48, 389. Hemmila, I., Applications of Fluorescence in Immunoassays, Wiley-Interscience, New York, 1991. Hemmila, I., Anal. Chem.. 1985, 57, 1676. Bador, R., Dechaud, H., Claustrat, F., and Desuzinges, C., Clin. Chem. (Winston-Salem, N.C.), 1987, 33, 48. Saarma, M., Jarvekiilg, L., Hemmila, I., Siitari, H., and Sinijarv, R., J. Virol. Methods, 1989, 23, 47. Hemmila, I., Holttinen, S . , Pettersson, K., and Lovgren, T., Clin. Chem. (Winston-Salem, N. C.), 1987, 33, 2281. Siitari, H., J. Virol. Methods, 1990, 28, 179. Hemmila, I.. in Fluorescence Spectroscopy, ed. Folfbeis, O., Springer, Heidelberg, 1992, in the press. Kononenko, L., Poluektov, N., and Nikonova, M., Zavod. Lab., 1964, 30, 779. Melenteva, E., Poluektov, N., and Kononenko, L., Zh. Anal. Khim., 1967, 22, 187. Yang. J.-H., Zhu, G.-Y., and Wu, B., Anal. Chim. Acta, 1987, 198, 287. Ci, Y.-X., and Lan, Z.-H., Analyst, 1988, 113, 1453. Ci, Y.-X., and Lan, Z.-H., Anal. Lett., 1988, 21, 1499. Yang, J.-H., Zhu, G.-Y., and Wang, H., Analyst, 1989, 114, 1417. Ci, Y.-X., and Lan, Z.-H., Anal. Chem., 1989,61, 1063. Yang, J.-H., Ren, X.-Zh., Zou, H.-B., and Shi, R.-P., Analyst, 1990. 115, 1505. Haddad, P., Talanta, 1977,24, 1. Yamada, S., Kano, K., and Ogawa, T . , Anal. Chim. Acra, 1982, 134, 21. Xu, Y.-Y.. Hemmila, I., Mukkala, V.-M., Holttinen, S., and Lovgren, T., Analyst, 1991, 116, 1155. Xu, Y.-Y., and Hemmila, I . , Anal. Chim. Acta, 1992, 256, 9. Xu, Y.-Y., and Hemmila, I., Talanta, 1992, in the press. Halverson, F., Brinen, J . , and Leto, J.. J. Chem. Phys., 1964, 41, 157. Halverson, F., Brinen, J . , and Leto, J . , J. Chem. Phys., 1964, 41, 2752. Xu, Y.-Y., Pettersson, K., Blomberg, K., Hemmila, I., Mikola, H., and Lovgren, T., unpublished work. Beinlich. C., Piper, J . , O’Neal, J . , and White, O., Clin. Chem. (Winston-Salem, N.C.), 1985, 31, 2014. Apter, D., Cacciatore, B., Alfthan, H., and Stenman, U.-H., J. Clin. Endocrinol. Metab., 1989, 68. 53. 1985, pp. 203-217. References 1 Soini. E . , and Hemmila, I., Clin. Chem. (Winston-Salem, N.C.), 1979, 25, 353. Paper 2100587E Received February 3, 1992 Accepted March 3, 1992