tecting amphetamines in urine. ... amphetamine assay, we used a non-denaturing detergent, ..... Effect of DOS on the pH dependence of the fluorescence.
CLIN. CHEM. 32/9, 1677-1681 (1986)
Homogeneous,MicelleQuenchingFluoroimmunoassay for Detecting Amphetaminesin Urine Clarke J. Halfman and DennIs W. Jay We developed a homogeneous fluoroimmunoassay for detecting amphetamines in urine. Only fluorescence intensity need be measured because the emission of non-proteinbound fluorescein-labeled amphetamine is preferentially quenched by detergent micelles. In a previously reported prototypeassay system for measuring gentamicin in serum
we used fluoresceinand dodecyl sulfate (Anal Chem 1985; 57:1928-30). We have foundthat favorablehydrophobicand (or) ionic character of the analyte and unfavorable polar and (or) ionic character of the fluor are important determinants of the desired interactions. An anionic detergent and fluorescein, therefore, should be appropnate for apolar or cationic analytes, such as gentamicin and amphetamines. A greater [H] at the anionic micelle surface is important for quenching emission from the fluor moiety. Millimolar concentrations of dodecyl sulfate rapidly denature immunoglobulin unless hapten is bound with sufficiently high affinity. Affinity was sufficiently high for the antibody used in the prototype gentamicin assay but not for the amphetamine antibody. Thus for the amphetamine assay, we used a non-denaturing detergent, dodecyl(oxyethylene)12 sulfate. The assay requires 30 j.L of specimen in 2 mL of total assay volume. Amphetamine (d-, dl-, and meth-), at a concentrationof 1 mg per liter of unne, is readily detected.
AddItional Keyphrases:methamphetamine gents on Immunoglobulins
.
effects of deter-
dodecyl(oxyethylene)12
sulfate
A homogeneous fluoroimmunoassay response is provided when a fluorescence property of bound labeled analyte is substantially different from that of free labeled analyte. Fluorescence intensity is the simplest and most sensitive property to measure, but usually is not considerably different for free and bound labeled analyte. Although polarization is generally greater for bound labeled analyte, and is therefore a useful response variable in immunoasaays for analytes of small molecular size (1), a general means for inducing an intensity difference between bound and free
labeled analyte would provide for a simpler, potentially more sensitive, less expensive assay. Three such means have been reported in the literature: conjugating antibody with a complementary acceptor dye quenchesemission,by energy transfer, specifically from bound labeled analyte (2); using an additional antibody to fluorescein quenches emission specifically from free labeled analyte, when steric hindrance from binding by analyte antibody prevents antifluor binding (3); and detergent micelles can preferentially quench emission from free labeled analyte (4). We demonstrated the last approach with an assay for measuring gentamicin in serum, using dodecyl sulfate Department, University ofHealth Sciences/TheChicaSchool, North Chicago, IL 60064. Received May 5, 1986; acceptedJune 24, 1986.
Pathology go Medical
(DDS) and fluorescein-labeled gentamicin.’ The general utility of the method was not established, but we postulated that the preferential quenching of emission from free labeled analyte was due to micelles interacting directly with the analyte moiety but not with the fluor moiety. Thus, if the analyte moiety were envelopedwithin the immunoglobulin binding site, then the fluor moiety would not interact with micelles, even though exposed to the bulk solvent, and emission would not be quenched. We suggested that the requisite interactions could be achieved by using a detergent and a labeling fluor with the same ionic charge (and opposite to that of the analyte, when charged). However, molecular features other than ionic charge are evidently responsible for the lack of interaction between fluorescein and DDS micelles, because other ionic fluors clearly interact with detergent micelles regardless of respective ionic charges (5). To evaluate the potential general utility of detergent micelles for providing a homogeneous fluoroimmunoassay response, we needed to elucidate the molecular features that inhibit the interaction of some fluors, such as fluorescein, with ionic detergents, and that would assure interaction of the analyte moiety with micelles. For this purpose we compared the interaction of various ionic fluors with a catiomc and an anionic detergent. Furthermore, we suspected that intensity might not always be quenched when labeled analyte interacted with micelles. Indeed, it was surprising that fluorescein emission was altered, because the emission of other fluors appeared to be altered by the incorporation of the fluors into the hydrophobic micelle interior (6-8). We therefore conducted studies to examine the mechanism of the micelle-induced quenching. DDS is known to denature proteins (9), including immunoglobulins (10). However, we recently reported that hapten binding prevented detergent denaturation of immunoglobulins so long as binding affinity for hapten was sufficiently high (11). The affinity of the immunoglobulin used in the prototype gentamicin assay (4) was apparently sufficiently high, but the lower-affinity amphetamine antiserum was rapidly denatured by DDS (11). For use with this amphetamine antiserum we therefore found a less-denaturing detergent, dodecyl(oxyethylene)12 sulfate [DD(OE)12S], which provided a homogeneous distinction between free and bound fluorescein-labeled methamphetamine.
Materials and Methods Chemicals. We used glass-distilled, de-ionized water. All other reagents were at least reagent-grade quality and were usedwithout further purification. The sodium salt of DDS and the bromide of dodecyl trimethyl ammonium (DDTA) were purchased from Aldrich Chemical Co., Milwaukee, WI; the sodium salt of DD(OE)12S was obtained from Henkel, ‘Nonstandard abbreviations: DD(OE),2S, dodecyl(oxyethylene),2 sulfate;DDS, dodecyl sulfate; DDTA, dodecyltrimethylammonium; FA, fluoresceun-labeled methamphetamine. CLINICALCHEMISTRY, Vol. 32, No. 9, 1986 1677
Inc., Fort Lee, NJ. Fluorescein was the “Kromex” product from J. T. Baker Chemical Co., Phillipsburg, NJ. Phosphate-buffered saline, containing 10 mmol of NaH2PO4, 150 mmol NaCl, and 15 Mxnolof DDS per liter, was adjusted to the desired pH (usually 7.4) with HC1 or NaOH. The low concentration of DDS maintained stability of solutions containing low concentrations of fluorophores. Borate-buffered saline contained 10 mmol of sodium borate, 150 mniol of NaCl, and 15 pmol of DDS per liter, and was adjusted to the desired pH as above. Sheep antiserum to amphetamine, a generous gift from Syva Co., Palo Alto, CA, was stored in aliquots at -80 #{176}C until needed. Preparation offluorescein-labeled methamphetamine. Fluorescein-labeled methamphetamine (FA), with a butylamine bridging group, was prepared from fluorescein isothiocyanate and methamphetamine, both from Sigma Chemical Co., St. Louis, MO, under reaction conditions described by Smith (12). The N-4-(aminobutyl)methamphetamine was synthesized by reacting methamphetamine with 4-(bromobutyl)phthaliniide (Aldrich) as described by Aoki and Kuroiwa (13). The purity of the isolated product was evaluated by thin-layer chromatography on precoated silica-gel F-254 plates (E. Merck, Westbury, NY) with a developing solvent of chloroformlmethanol (70/30 by vol). A single, well-defined fluorescent spot was observed under long-wavelength ultraviolet illumination (RF = 0.12; the RF of fluorescein isothiocyanate under those conditions was 0.32). We based FA concentrations on absorbance measurements at 492 nm and the molar absorptivity of 9.0 x iO mol’ cmi, given by the manufacturer, of fluorescein at this wavelength. Fluorescence measurements. Fluorescence properties were measured with a dual-channel, Model 4000 SLM spectrofluoropolarimeter (SLM Instruments, Urbana, IL) as described previously (14). We used an excitation wavelength of 494 nm and matched Turner 2A12 filters (transmission above 500 nm), purchased from Sequoia-Turner Corp., Mountain View, CA, in the detection beams for measurements of fluorescein. The appropriate excitation wavelength and matched cutoff filters were used for measuring the other fluorescent dyes. Glann-Thompson polarizers, supplied with the instrument, were used in the excitation and emission beams when necessary. Vertical (V) and horizontal (H) emission components were measured simultaneously on the two individual detection channels with vertically polarized excitation. We normalized the signal strengths of the two detection channels by setting V = H, with horizontally polarized excitation for polarization determinations. When V - Hwasusedtomonitorfluorbinding(15, 16),wesetV = H, with vertically polarized excitation for the appropriate fluor reference solution. Design of the assay. The anti-amphetamine serum was characterized to optimize the assay response. We determined binding-site concentration and affinity by titrating two dilutions of antiserum with FA (17) and measuring V H as the response to binding (15, 16). The titration data were fit by nonlinear, least-squares regression analysis (18) with binding-site concentration and affinity for FA as the variables. Binding-site concentration of the antiserum was 51 (SD 2) pmoIJL and the equilibrium dissociation constant was 7.5 (SD 0.5) ninoiiL. The equilibrium dissociation constant for methamphetamine, 10.8 (SD 1.4) nmol/L, was determined by fitting data from competitive binding between the drug and FA (18). The concentration of antibodybinding sites (40 nmol/L) and of FA (52 nmol/L) for the assay were chosen to provide a response with a maximum initial .
1678
CLINICAL
CHEMISTRY,
Vol. 32, No.9, 1986
relative slope and a midpoint at a drug concentration of 1 mgfL, with use of 30 L of specimen and a total assay volume of 2.0 mL (19). Assays were conducted by adding 30 ,uL of standards or samples to 1.95 mL of FA in phosphatebuffered saline, pH 6.5, containing 50 mmol of DD(OE)12S per liter; initial intensity (ii) was measured several minutes later. Final intensity (12) was measured several minutes after adding 20 L of a 13-folddilution of antiserum. Assay response (M was Ii - ‘2.
Resufts and Discussion Detergent-Dye Interactions We gained information about the molecular features of the co-reactants responsible for the requisite interactions by examining the influence of micelles of a cationic detergent, DDTA, and of an anionic detergent, DDS, on the fluorescence polarization of several ionic fluorescent dyes (see Figure 1). As shown in Table 1, all of the fluors interacted to some extent with each detergent above the critical micelle concentration. However, the anionic fluors interacted more strongly with the cationic micelles and the cationic fluors interacted more strongly with the anionic micelles. Rhodamine B interacted strongly with both detergents. With DDS, fluorescein interacted the most weakly; with DDTA, the pyronin Y interaction was the weakest. The stronger interaction between other like-charged fluors and detergent was apparently due to the presence of apolar groups, i.e., the benzoyl ester moiety of methyl fluorescein and of rhodamine 6G. Indeed, fluorescein interacted with DDS micelles at pH values less than 5.5 when the benzoate moiety was neutralized-behavior consistent with results from previously reported studies of other dye-detergent systems. The solubiization of apolar fluors such as pyrene by detergent micelles clearly demonstrates that apolar compounds, or groups, are readily incorporated into the hydrophobic micelle interior (7,20). The presence of a polar group at one end of a co-reacting compound would limit the extent of penetration of the apolar moiety because the energetically favorable location of the polar group would be at the micelle surface exposed to the polar, bulk aqueous solvent, as would be the polar head group of the detergent monomers composing the micelle. When the detergent and the co-reacting species are ionic and of the same charge, the depth of H
0
..--.-
‘okokoJ
OH
LOkOIOJ OEt
methyl Fluorescein
Fluorescein
,N
0
NH jMs),
(M.)15JJ6JiI.)2
Acridine Orange
Pyronin V
H
N
()a OEt
Rhodamins 60
FIg. 1. Structures of fluorescent dyes
0
0 0 0 0
(t)2
g#{176}c
Rhodamine B
Table 1. interaction of ionic Fluorescent Dyes with a Cationic and an Anionic Aikyi Chain Detergent in PhosphateBuffered Saline (pH 7.4) DDS concn,
Fluorescentdye (5 nmoiIL)
Fluorescein Methylfluorescein Acridineorange Pyronin V Rhodamine6G
0.001: polarization 0.014
a
mmel/L
DDTA cencn,’
20 Intensity9
0.016 0.030 0.027
0.025 0.070
1.00
20
0.001:
polarization
intensity9
Polarization
0.020
0.80
0.095
0.058 0.020 0.097 0.028 0.107 0.031 0.165 0.028 0.165 0.070 and 5 mmoVL,respectively.
0.60
0.095 0.080 0.034 0.130 0,180
Polarization 0.018
0.47
0.53 0.53 0.35 3.0
Rhodamine B The critical micelle concentrations for DDS and DDTA are approximately I 9Relative
mmel/L
0.65
1.10 0.15 1.67
to the intensityIn 0.001 mmol/Ldetergent.
penetration of an apolar moiety of the co-reacting molecule is further limited by electrostatic repulsion.When the ionic charge of the detergent and of the co-reacting species is opposite,then electrostaticattraction addsto the interaction energy and mixed micelles may even form at a detergent concentration less than the critical micelle concentration (21). Interestingly, fluorescein did not interact with DDS micelles when the tricyclic dihydroxy xanthene group was neutralized (plC0 = 6.2). The high polarity of the oxygen heteroatom, rather than the rigid, planar structure of the zanthene group, probably prevented incorporation of this neutral group into the hydrophobic micelle interior. This corjecture was substantiated by the observation that pyronin Y did not interact with DDTA micelles, although acridine orange did (Table 1). In light of this behavior, we concluded that a molecule would be inhibited from interacting with micelles only if like ionic charges and (or) polar groups were appropriately located to prevent incorporation of apolar groups into the hydrophobic micelle interior. Apolar groups of such a molecule would not be incorporated into the micelle interior, even when ionic charges are opposite to the micelle charge, but the molecule would be “adsorbed” at the micelle surface by electrostatic attraction, just as counter-ionsare concentrated at the micelle surface (22, 23). Fluorescein and pyronin Y, as well as rhodamine B, were therefore probably adsorbed at the surface of oppositely charged micelles. Apparently, this was also the mode of interaction of genta. micin with DDS micelles in the prototype assay (4). Interaction of the other fluors (Table 1) involved incorporation of apolar portions into the micelle interior. The lack of interaction of fluorescein with DDS micelles was accounted for by several molecular features: the carboxylate of the benzoate moiety, the ionized quinoid group of the dihydroxy xanthene tricyclic ring system, and the oxygen heteroatom of the middle ring preventing interaction, even when the quinoid group was neutralized. The combination of fluorescein as labeling fluor and an anionic detergent such as DDS would be well suited for use in assay systems for any apolar or cationic analyte. Although a cationic detergent such as dodecyl trimethyl alnmonium would be suitable in assays for any apolar or anionic analyte, a candidate fluor is more problematic. Pyronin Y did not interact with micelles of a cationic alkyl chain detergent, but reactive derivatives for coijugation to analyte are not available. Reactive derivatives of rhodamine B are available, but this fluor interacted with DDTA micelles (Table 1), apparently by electrostatic attraction of
the benzoate group. Possibly, when conjugated to analyte and bound to antibody, steric hindrance may prevent the benzoate group of rhodamine B from interacting with micelles. Otherwise, synthesis of the choline ester should provide a suitable labeling fluor. Drug Detection Detection of abused drugs in urine is a useful application of the method, most drugs of abuse, except the barbiturates, being cationic amino compounds; thus fluorescein and an anionic detergent (e.g., DDS) would be appropriate components of an assay system. We developed an assay system for amphetamines first. The fluorescein-isothiocyanate adduct of butylamine-bridged methamphetamine (FA) interacted with DDS micelles in phosphate-buffered saline (pH 7.4) as expected. Evidently, the amphetamine moiety was responsible for the interaction, because fluorescein itself did not interact. The interaction quenched fluorescein emission by 75%. The mechanism for the micelle-induced quenching was suggested by the effect of DDS micelles on the pH dependence of emission from the fluor moiety (Figure 2). The titration of FA in the absence of DDS (curve a) was essentially identical to that of fluorescein (plC0 of 6.2); DDS micelles shifted the plC0to 8.0 (curve b). The concentration of counter-ions, including H, may be several orders of magnitude greater at the micelle surface than in the bulk solvent (22, 23). The results from the titrations (Figure 2) therefore suggested that the micelle-induced quenching was 1.0
0.8
‘-
a
-
0.6
0.4
0.2
0.0 5
6
8
9
10
pH
Fig. 2. Effect of DOS on the pH dependence of the fluorescence intensityof FA FA. 10 nmol/L, in phosphate- or borate-butteredsaline, with (b) or without (a) DDS, 10 mmol/L,was titratedto the indicatedpHandthe intensity was measured. The detergent micellesshiftedthe pK.of the fluorophorefrom about6.2 toabout
8.0 CLINICALCHEMISTRY, Vol. 32, No. 9, 1986 1679
dueto protonationof the fluorophoricgroup because of a 60fold greater H concentrationat the micelle surface. By this mechanism, the assay response magnitude is obviously pH dependent. Assuming that the intensity of antibody-bound FA is equivalent to the intensity in the absence of DDS, the absolute response magnitude would be greatest at pH 7.1. The magnitude of the more pertinent relative response would be greater the lower the pH. In practice, the minimum pH to be used would be limited by the signal strength needed to avoid appreciable values for specimen blanks. DS was unsuitable for the amphetamine assay because the detergent rapidly denatured the antibody used, haptenbinding ability being rapidly lost when DDS was added (final concentration 10 mmol/L) to antibody-bound FA. We reported earlier that bound hapten could prevent DDS denaturation of immunoglobulin so long as affinity was sufficiently great, or dissociation rate was sufficiently slow (11). The binding of fluorescein-labeled thyroxin to thyroxin antibody, with a dissociation half.time of 1.25 h, protected from DDS denaturation. Evidently, the affinity of the antibody used in the prototype gentamicin assay (4) was also sufficiently great. However, the dissociation rate (t#{189} = 75s) of the anti-amphetamine immunoglobulin we usedwas too rapid to afford adequate protection from DDS denaturation. Of several other anionic detergents evaluated for their non-denaturing ability, DD(OE)12S seemed most suitable. Micelles of this detergent interacted with FA and consequently quenched emission (Figure 3). The effectof micelles of this detergent on the pH-dependence of FA intensity was similar to that of DDS (Figure 2) except that the plC0was shifted only 1.3 pH units, representing a somewhat lesser increase in H concentration, about 20-fold, at the micelle surface.
This detergent, however, was less than ideal because its inicelles also interacted with antibody-bound FA. We suspected micelle interaction with the fluorescein moiety and indeed demonstrated interaction of DD(OE)12S micelles
(Figure 4). The micelle interaction inthe pK0 of the fluorogenic group by about 1.0 pH unit. Apparently, the neutralized dihydroxy xanthene group favorably partitioned within the (oxyethylene) micelle shell by about ninefold more than in the bulk aqueous
with
fluorescein
0.20
0.15 0 0
0.10
0.50 0
0 0.
0.05
0.00 pH
Ag. 4. Effect of DD(OE)12S on the pH dependence of fluorescent
propertiesof fluorescein Fiucreeceln,10 nmol.t. with or without DD(OE)S, 50 mmol/t.,was titrated In phosphate-butteredsaline or In borate. Curves a and d illustrateintensityand polarization,respectively,In the absenceof detergent Curveeband c illustrate intensity and polarization,respectively,inthe presence of the detergent
solvent. We did not observe this behavior with the completely hydrophobic DDS micelles. Evidently, only the weakly fluorescent, neutralized form of the dihydroxy xanthene group was incorporated into the polar shell of the DD(OE)12S micelles; otherwise, polarization would have been increased at higher pH. Polarization was increased only at lower pH, where the neutralized, micelle-incorporated form contributed substantially to the total emission. Even though DD(OE)12S micelles interacted with the fluorescein moiety of FA bound to amphetamine antibody, the interaction was stronger with free FA. The relatively greater quenching of free FA emission by DD(OE)12S micelles increased with decreasing pH; at a pH of 6.5 or less, the intensity offree FA was less than 60% of bound FA. This extent of differential quenching by DD(OE)S micelles was sufficient to provide a homogeneous distinction between free and bound FAin an immunoassay. A typical response curve for methamphetamine is shown in Figure 5. Analysis of 50 different urine specimens, to which we added1 mg of d-, d,l-, or methamphetamine per liter, demonstrated an adequate
creased
2.5
2.0 1.00
M
0.75
1.0
0 C
.0.50
C
=
0 0.
0.5 0.25
0.0
0.0 0.00 0.01
0.10
1.00
[DD(OE),2SJ,
10.00
100.00
mmol/L
Fig. 3. fluorescent titration of FA with DD(OE)12S
V&iousosnoentrallons of DO(OE),2Swere addedto FA. 10 nmo4&.In a solution contaIning10 mmolof sodium phosphateand 1 mel of Nsa per liter at pH 6.5. Increased polarization ( and decreasedintensity(a) were aseodatedwith kieraction with micelles
1080
CUNICAL
CHEMISTRY,
VoL 32, No.9, 1986
0.5
1.0
1.5
2.0
2.5
3.0
(Methamphetan,ine), mg/L
Fig. 5. Response curve for methamphetamlne fluoroimmunoassay Methamphetaminestandards (30 ML) were added to FA. 52 nmo and DO(#{176}E)12S, 50 mmol/L in phosphate-butteredsalIne(pH 6.5) and initialIntensities were measured.FInalintensitleewere measured 15 mm after adding antiarv#{231}hetamine to a final bindIng-sIteconcentrationof 40 nmot&Ina totalvolumeof 2.0 mL Assayresponse( Is the differencebetweenfinal and InitialIntensity measurements
Table 2. PrecisIon of the Micelie Quenching Fluoroimmunoassay for DetectIng Amphetamines In UrIne Amph.t.mine
n
Response (‘O msanaSD 2.20 ± 0.20 1.24 ± 0.14 1.00 ± 0.08
None 48 d,I-Arnphetamine 36 Methamphetamine 8 d-Amphetamine 4 1.25 ± 0.04 a Drugadded to urine,1 mg/L final concentration.
drugcencn, mg/L (±1 SD)
0.05 0.20 0,20
intensity difference from that for unsupplemented controls (Table 2). Sensitivity (detection limit) was ± 0.05 mgfL, and precision at 1 mg/L was ±0.2 mg/L. The antiserwn exhibited highest affinity for methamphetamine, in contrast to the EMIT assay (Syva Co.) with the same antiserum, which exhibits the greatest response for 1amphetamine. The different relative affinities probably resulted from the different labels used in each assay system. At the pH of 6.5 used to obtain a reasonable magnitude of relative response, fluorescein emission was only a fractionof its value at higher pH, so that intensities of sample blanks were about 20% of the test signal. Conducting the assay at a higher pH would substantially minimize the relative signal strength of sample blanks, thus obviating their measurement and subtraction. Other quenching ions, such as Cu2, are also concentrated at the surface of oppositely charged micelles (22,23) and we are currently investigating whether their use will permit conducting the assay at higher pH. This paperispresented in partial fulfillment ofD. W. J.’s requirementsfor the Ph.D. References
1. Jolley ME, Stroupe SD, Wang CJ, Panas HN, Keegan CL. Polarization fluoroimmunoassay for aminoglycosides in serum. Clin Chem 1981;27:1190-7. 2. Ullman EF, Schwarzber M, Rubenstein KE. Fluorescent excitation energy transfer ixnmunoassay. J Biol Chem 1976;25l:4172-8. 3. Zuk RF, Rowley GL, Uliman EF. Fluorescenceprotection immunoassay: a new homogeneous assay technique. Clin Chem 197925:1554-60. 4. Halfman CJ, Wong FCL, Jay DW. Solventperturbation fluorescenceimmunoassay technique. Anal Chem 1985;57:1928-30. 5. De Vendittis E, Palumbo G, Parlato G, BocchiniV. A fluorimetnc method for the estimation of the critical micelle concentration of surfactants. Anal Biochem 1981;115:278-86. 6. Hautala RH, Schore NE, Turro NJ. A novel fluorescent probe. Use of time-correlated fluorescence to explore the properties of micelle-forming detergents. J Am Chem Soc 1973;95:5508-14.
7. Ohyashiki T, Mohri T. Fluorometric analysis of micelle formation of sodium dodecyl sulfate. Chem Pharm Bull 1978;26:3161-6. 8. Wolff T. The solvent dependent fluorescence quantum yield of acridine as a probe for critical micelle concentrations. J Colloid Interface Sci 1981;83658-60. 9. Reynolds JA, Tanford C. The gross conformation of proteindodecyl sulfate complexes. J Biol Chem 1970;245:5161-5. 10. Nelson CA. The binding of detergents to proteins. I. The maximum amount of dodecyl sulfate bound to proteins and the resistance to binding of several proteins. J Biol Chem 1971;246:3895-901. 11. Halfman CJ, Dowe R, Jay DW, SchneiderAS. The effect of dodecyl sulfate on immunoglobulinhapten binding. J Mol Immunol 1986; in press. 12. Smith DS. Enhancement fluoroimmunoasaay of thyroxin. FEBS Lett 1977;77:25-7. 13. Aoki K, Kuroiwa Y. Enzyme immunoassay for methainphetamine. J Pharmacobio-Dyn 1983;6:33-8. 14. Halfman CJ, Wong FCL, Schneider AS. Direct measurement of fluorescence polarization or anisotropy. Anal Chem 1983;55:14324. 15. Roewelt PM, Capon AJ, Seitz WR. Selective analysis of binary fluorophore mixtures by fluorescence polarization. Anal Chem 1980;52:769-70. 16. Halfman CJ, Wong FCL, Schneider AS. Directly measured signals as response variables in fluorescencepolarization ligand binding assays. Anal Chem 1984;56:1648-50. 17. Halfman CJ, Nishida T. Method for measuring the binding of small molecules to proteins from binding-induced alterations of physical-chemical properties. Biochemistry 1972;11:3493-8. 18. Wong FCL. An improved response variable for fluorescence polarization immunoassay and optimization of reactant concentrations. Ph.D. thesis, University of Health Sciencesfflie Chicago Medical School, 1984. 19. Halfman CJ, Schneider AS. Optimization of reactant concentrationsfor maximizing sensitivities of competitive immunoassays, Anal Chem 1981;53:654-8.
20. Singh H, Hinze WL. Micellar enhanced spectrofluorimetric methods: application to the determination of pyrene. Anal Lett 1982;15:221-43. 21. Sato H, Kawasaki M, Kasatani K. Energy transfer between rhodamine 6G and pinacyanol enhanced with sodium dodecyl sulfate in the premicellar region. Formation of dye-rich induced micelles. J Phys Chem 1983;87:3759-69. 22. Foreman TIC, Sobol WM, Whitten DG. The importance of hydrophobic-hydrophilic factors in binding of charged substrates to micelles: the use of extramicellar probe luniinescence to monitor association of cations to the micelle. J Am Chem Soc 1981;103:5333-6. 23. Abulin E, Lissi E, Bianchi N, Miola L, Quina FH. Quenching of aromatic hydrocarbonfluorescence by counterions in aqueousmicellar solution. Relationship to ion exchange. J Phys Chem 1983;87:5166-72.
CLINICALCHEMISTRY,
Vol. 32, No. 9, 1986 1681
