Biosensors& Bioekctronics 6 ( 199 1) 507-5 16

Pharmacological specificity of a nicotinic acetylcholine receptor optical sensor Kim R. Rogers, Mohyee Department

of Pharmacology

and Experimental Therapeutics, MD 21201. USA

Darrel E. Menking, Biotechnology

E. Eldefrawi*

Roy G. Thompson

Division. US Army Research Development

University

of Maryland,

Baltimore.

& James J. Valdes

& Engineering

Center, Edgewood, MD 21010. USA

(Received 7 May 1990; revised version received 10 October 1990; accepted 1I October 1990)

Abstract: The pharmacological specificity of a nicotinic acetylcholine receptor (nAChR) optical biosensor was investigated using three fluorescein isothiocyanate (FITC)-tagged neurotoxic peptides that vary in the reversibility of their receptor inhibition: a-bungarotoxin (a-BGT). a-Naja toxin (a-NT), and aconotoxin (GI) (a-CNTX). Kinetic analysis of the time course of binding of FITC-neurotoxins to the nAChR-coated fiber gave association rate constants (k,,) of 8.4 X 106 M-’ min-’ for FITCu-BGT, 6-O X 106 M-’ min-’ for FITC-oNT and 1.4 X 106 M-’ min-’ for FITC-a-CNTX. The dissociation rate constants (k-i) for the three neurotoxins were 7.9 X 10e3 min-‘. 4.8 X IO-’ mitt-’ and 8.0 X 10-i min-’ for FIT&z-BGT, FITCa-NT and FITC-o-CNTX respectively. The equilibrium dissociation constant (&) values for the three toxins, calculated from these rare constants, were similar to published values obtained from tissue responses or ligand binding assays. The optical signal generated by FITC-a-NT binding to the nAChR-coated fiber was effectively quenched by agonists and antagonists of the nAChR but not by most of the tested agonists and antagonists of muscarinic cholinergic, adrenergic. glutamatergic. serotonergic. dopaminergic or GABAergic receptors. Interestingly. 5-hydroxytryptamine, haloperidol and (+ )cis-methyldioxolane gave significant inhibition of FITCa-NT binding to the immobilized receptor. Equilibrium constants of inhibition (Ki) for d-tubocurarine (d-TC) and carbamylcholine (carb) were determined from competition studies using FITCar-CNTX. FIT&z-NT or FITC-a-BGT as probes for receptor occupancy. When the more reversible probe FITCu-CNTX was used, the Ki value ford-TC was an order of magnitude lower than those determined using the less reversible probes. K, values for carb. however, were independent of the FITC-toxin probe used. Keywords: nicotinic tagged toxins.

*To whom correspondence 0956-5663/91/$03.50

acetylcholine

receptor, fiber optic biosensor,

fluoresein-

should be addressed.

Q 1991 Elsevier Science Publishers Ltd.

507

Kim R. Rogers et al. INTRODUCTION Biosensors are analytical devices that utilize a biological material (e.g. enzymes, antibodies, receptors or whole tissues) immobilized on a physical transducer (e.g. electrode, electronic chip, or optical fiber) for the detection of specific chemicals (Rechnitz, 1987; Sepaniak et al., 1989; Buch & Rechnitz, 1989; Guilbault & Luongm, 1989; Scheller et al., 1989; Wise, 1989). Excellent examples include the variety of biosensors that have been developed for detection of glucose (Get-net ef al., 1989; Gotoh ef al., 1989~; Hsuie ef al.. 1989) and immunosensors which utilize antibodies for the detection of specific antigens (Hirshfeld & Block, 1984; Andrade et al., 1985; Davis & Leary, 1989; Gotoh et al., 1989b; Rishpon &Rosen, 1989). These immunosensors depend largely on the high affinity recognition that antibodies have for their substrates. Other biological macromolecules that depend on high affinity recognition are neurotransmitter and hormone receptors. These are the natural biosensors of our bodies, whose physiological role dictates that they recognize and bind specific chemical signals (i.e. hormones and neurotransmitters) with high affinity. Investigations of receptor-based biosensors has been limited to the nicotinic acetylcholine receptor (nAChR) of electric organs (Eldefrawi er al., 1988; Gotoh et al., 1987; Taylor et al.. 1988; Rogers et al., 1989). Although receptors typically have high affinity for their specific chemical signals (i.e. neurotransmitters or hormones), many of them have even higher affinities for natural or synthetic antagonists. For example, muscarinic acetylcholine receptors (mAChR) have lower affinities for acetylcholine and other agonists (i.e. PM Kd values) than for antagonists such as quinuclidinyl benzilate (QNB) and atropine which have Kd values in the nM to PM range (Birdsall & Hulme, 1976). Further, the nAChR receptor has a dissociation constant in the PM range for ACh and agonists but nM to PM dissociation constants for the antagonist a-neurotoxins of snake venoms (e.g. a-bungarotoxin (a-BGT) and a-Nuju toxin (a-NT)) (Albuquerque et al., 1979). Therefore, specific binding of QNB and a-BGT are commonly accepted to identify mAChR and nAChR, respectively. Nevertheless, in addition to their pharmacological targets. drugs and toxins bind nonspecifically to glass surfaces (Cuatrecases & Hollenberg, 1975) plasma proteins (Goldstein 508

Biosensor.9& Bioelectronics6 ( 199I ) 507-5 I6 etul., 1969) and other macromolecules in tissue preparations. The latter is generally termed nonspecific binding and is regularly encountered in radioactive ligand binding assays (O’Brien, 1986). Pharmacological specificity of a binding event is an important criterion for a successful receptor biosensor. Specific binding of receptor ligands should exhibit kinetics similar to their in situ kinetics. In addition, binding should exhibit a pharmacological profile which matches that of the in situ receptor response to various drugs. This appears to be the case for the nAChR optical sensor which bound fluorescein isothiocyanate (FITC)-labeled a-BGT (FIT&x-BGT) with high affinity. Further, the binding of FITCa-BGT was inhibited by agonists and antagonists of the nAChR (Rogers et al., 1989). Because binding of a-BGT to the nAChR is semi-irreversible while a beneficial property for a biosensor is its reusability, two other FITC-tagged competitive inhibitors were evaluated in the present study: the more reversible cobraNuju nuju venom a-NT and the completely reversible venom of the snail Conus geogruphus, conotoxin GI (CNTX GI) (Wu & Narahashi, 1988). The binding and dissociation kinetics of FITCaBGT, FIT&x-NT and FITC-a-CNTX were compared using the nAChR immobilized on optic fibers. The nAChR has different affinities for the three toxins, in the PM range for a-CNTX (Hashimoto etul., 1985) compared to its high affinity for a-NT and a-BGT (Colquhoun & Rang, 1976). In addition, the effect of cholinergic and noncholinergic ligands on the optical signal generated by binding of FITC-a-NT to nAChR coated fibers was investigated as a measure of the pharmacological specificity of the response.

MATERIALS

AND METHODS

Preparation of fluorescein-labeled toxins

The neurotoxins a-BGT and a-NT were obtained from Ventoxin Laboratories (Frederick, MD). QCNTX was a gift from Dr A. Galdes of the BOC Group Inc. (New Jersey). Coupling of FITC to these neurotoxins was performed as previously described by Rogers ef al. (l989), with minor modifications for FITC-a-CNTX. Toxins (2 mg) were reacted with 1 mg FITC on celite (Sigma) in 1 ml of 50 mM bicarbonate buffer at pH 9.5 for

Biosertsors& Bioelectronics6 ( 199 1) 507-5 16

15 min. The celite was removed by centrifugation and the supernatant for the FIT&r-BGT and FITCa-NT loaded onto a Sephadex G-25 column (25 cm X I.1 cm). For a-CNTX the supematant was loaded onto a Sephadex G-15 column (25 cm X I.1 cm). The columns were developed with 5 mM ammonium acetate, pH 5.8 and the void fractions pooled and lyophilized. The FITCa-CNTX was resuspended in phosphate-buffered saline (PBS; 154 mM NaCl in 10 mM sodium phosphate buffer, pH 7.4) at 1.8 mM and stored at -80°C until use. The FIT&r-NT and FITC-oBGT were suspended in 50 mM ammonium acetate pH 5.8, loaded onto Whatman CM-52 columns (10 cm X 1.5 cm) and eluted with the same buffer. The first peak of fluorescence was pooled for FITCa-NT and the second pooled for FITCa-BGT. The pooled fractions were then lyophilized, suspended in PBS buffer and assayed for biological activity as previously described (Rogers er al., 1989). Purification and immobilization of receptor protein

The nAChR was purified as previously described (Rogers et al.,1989). Purified nAChR was noncovalently immobilized on the quartz fibers (60 mm X 1 mm diameter, non-clad) during a 30 min incubation at 4°C in a solution, which contained 50 ,ug ml -’ nAChR protein in 15 mM each of HEPES and Citrate. pH 3.5. After immobilization, excess receptor was removed from the fiber by a 1 min incubation in 5 ml of PBS. Apparatus

All experiments were carried out using a fluorimeter designed and built at Ord Inc. (North Salem, NH). The fiber optic evanescent fluorosensor apparatus configuration has been previously reported (Glass etal., 1987; Rogers etal., 1989). Components of this instrument include a 10-W Welch Allyn quartz halogen lamp, a Hamamatsu S-1087 silicon detector, an Ismatec fixed speed peristaltic pump, and a Pharmacia strip chart recorder. The nonclad quartz fibers, 1 mm in diameter with polished ends, were obtained from Ord Inc. The fiber optic evanescent fluorosensor makes use of the evanescent wave effect by exciting a fluorophore just outside the waveguide boundary

PharmacologicalspeciJicityof an optical sensor

(excitation center wavelength of 485 nm; full width at half maximum (FWHM) 20 nm). A portion of the resultant fluorophore emission, then becomes trapped in the waveguide, is transmitted back up the fiber and is detected after transmission through 510 nm long pass and center wavelength 530 nm, FWHM 30 nm filters. The flow cell allows the center 47 mm of a fiber 60 mm long to be immersed in 46~1, which was exchanged every 12 s. Fluorescence measurements

After immobilization of the nAChR the fibers were placed in the instrument and perfused with PBS, pH 7.4 containing bovine serum albumin (BSA) (0.1 mg ml-‘) and competing ligands as specified in the Results section. After pretreatment, the fibers were treated with 5 nM FITC-a-BGT, 5 or 10 nM FITC-a-NT or 1 ,UM FITC-a-CNTX in PBS containing BSA (0.1 mg ml-‘) and competing ligands as specified. Between experiments, the flow cell was washed with 1% SDS for 2 min followed by PBS for 10 min. Kinetic analysis Association kinetics

Kinetic analysis of the receptor-specitic binding of FITC-labeled toxins to the nAChR-coated fibers was performed as described by Franklin and Potter (1972). A feature of this model, as applied to the binding of a-BGT to the nAChR, described by the equation a-BGT + receptor +I -

k

k-1

a-BGT-receptor complex

is the assumption that k+, > k-1. In order to apply this model to the observed interactions between FITC-labeled toxins and the immobilized nAChR on the optical fiber, the following assumptions were made: (1) Since the flow rate was 0.2 ml min-’ and the total assay time was 10 min, the total assay volume was 2 ml. (2) The initial concentration of active toxin is defined in terms of picomol per 2 ml assay volume (i.e. 10 pmol per assay for FITCaBGT). (3) The initial concentration of receptor binding sites is determined from the total available sites

Kim R. Rogers et

Biosensors& Bioelectronics6 (1991) 507-5 16

al.

immobilized on the fiber (i.e. 4 pmol per assay; Rogers &al., 1989) and the assay volume. (4) The amount of toxin bound at time t is from the determined receptor-specific fluorescence. The relative fluorescence yield (i.e. fluorescent units per molarity unit) measured using a Gilford Fluoro IV spectrofluorometer, differed for each of the FITC-conjugated toxins. Thus, the molar fluorescence yield for each of the labeled toxins was adjusted relative to the fluorescence yield reported for FIT&z-BGT (Rogers et al., 1989). (5) The time course data were collected for the first 10 min of the reaction beginning after the first 14 s (i.e. the time required for the flow cell to till with labeled toxin). Dissociation kinetics

Dissociation kinetics were assumed to proceed as a first-order decay reaction and k _ 1was determined from the slope of the logarithm of the binding versus time.

RESULTS FITC-labeled a-BGT, a-NT and a-CNTX bound nonspecifically to quartz fibers as measured by evanescent fluorescence (Fig. 1, upper traces).

However, addition of BSA (0.1 mg/ml) to the medium, eliminated the nonspecific binding of the three labeled toxins (Fig. 1, lower traces). Relative fluorescence of equimolar solutions of each of these FITC-conjugates were lOO%, 67% and 3% for FIT&x-BGT, FIT&z-NT and FITCa-CNTX, respectively. The concentration of labeled toxin that was necessary to yield a fluorescence signal in the optical biosensor was a reflection of the relative affinity of the toxin to the immobilized nAChR, the number of fluorescein molecules conjugated to the toxin and the quantum efficiency of the conjugate. Because of differences in concentration necessary for fluorescent signal measurement, the binding time courses for fluorescently labeled a-BGT (5 nM), a-NT (10 nM) and a-CNTX (1 PM) to the immobilized nAChR appeared to be markedly different (Fig. 2). Nevertheless, kinetic analysis of the time course data yielded similar values for association rate constants (Table 1). Dissociation kinetics of the FITC-labeled toxins also appeared markedly different from each other (Fig. 2). In contrast to the association kinetics, however, analysis of the dissociation time courses for the FITC-labeled toxins yielded k-1 values which differed from each other by an order of magnitude (Table 1). Equilibrium dissociation constants were calculated from the

100 mV

I 5 min.

1

BSA (0.1 mg/ml)

FITC-a-BGT (5 lw

FITC-a-NT (10 ti)

l-

BSA (0.1 q/ml)

FITC-a-CNTX (1 ph4

Fig. I. The upper traces show nonspectjic binding of the FlTC-labeled toxins to untreated quartz fibers. The lower traces, which appear as horozontal lines. demonstrated that this binding is eliminated by addition of 01 mg ml-’ BSA. 510

Biosensors & Bioelectronics 6 (1991) 507-5 16

Pharmacological spectficiq of an optical sensor

1

DISSOCIATION KINETICS

ASSOCIATION KINETICS

PITC-a-IICT PK-a-CNTX

FITC-a-NT

c RTC-WCNTX 100 mV MXNT)

I

Fig. 2. Comparative association and dissociation time courses for binding of FITC-labeled toxins to nAChR-coated fibers FITC-toxin concentrations were 5 nM: 10 nu and 1 pu for FITC-a-BGT FlTC-a-NT and FITC-a-CNTX. respectively in PBS + 0. I mg ml-’ BSA. After a 30 min treatment with the labeled toxins. dissociation time courses were initiated by petjiising thef7ow cel1.swith PBS + BSA in the absence of labeled toxin. TABLE 1

Kinetic parameters Toxin FITC-o-BGT FIT&z-NT FIT&z-CNTX

for binding

of fluorescently k+t(h4-t mitt-‘) 8.4 x 106 6.0 x 106 1.4 x 106

labeled toxins to the nAChR-coated k_t(min-t) 7.9 x IO-’ 4.8 x 1O-2 8.0 x 10-l

fibers K%(M) 9.4 x lo-‘0 7.3 x 10-9 5.7 x 10-r

“Kd values were determined using the relationship K, = k-,/k,,. The Kd value reported in bioassays of a-CNTX is 2 X 10m7M (Hashimoto et al., 1985) for the binding of a-BGT to rat muscle homogenates is lo-r0 M (Colquhoun & Rang, 1976) and for binding of a-NT to Torpedo membranes is 4 X 10m9M (Johnson & Taylor, 1982). relationship Kd = k_,lk+, (Table 1). Kd values are the reciprocal of affinity constants K, and are usually used as a measure of affinity. The relative differences in equilibrium constant values for the three toxins was determined primarily by differences in their dissociation rather than association rate constants. Specific binding of FITC-a-NT, as measured by the initial rates of fluorescence change (Fig. 3A) and maximum changes in fluorescence at equilibrium (Fig. 3B), were dependent on the concentration of FITCa-NT. Dose-response curves for both initial rates and maximum binding were similar in shape, increasing linearly with toxin concentration to about 10 nM then saturating (Fig. 3). The effects of various drugs on the optical

signal generated by binding of FITC-cz-NT to the nAChR-coated optical fiber was investigated to determine the pharmacological specificity and sensitivity of the optical biosensor. As shown in Table 2, the signal was inhibited significantly by the nicotinic cholinergic ligands. 5-Hydroxytryptamine (5HT) and haloperidol also gave significant inhibition, but GABAergic, glutamatergic and adrenergic ligands had no effect. The muscarinic cholinergic agonist cis-methyldioxolane also caused some inhibition. At lower concentrations, only the nicotinic ligands gave significant inhibition of the optical signal (Table 2). Equilibrium inhibition constants Ki for each of the toxins, using their respective FITC-labeled conjugates as probes, were 6.0 and 10.0 nM for a-BGT and a-NT, respectively and 1.5~~ for

Kim R. Rogers et

OY 0

al.

Biosensors & Bioelectronics 6

’

.

5

Fig. 3. The effect of FLY-a-NT initial

rate of jluorescence

’

’

15

10

FITC-NT

’

(nM) concentration

on (A) the

increase, or (B) the maximum

signal at 10 min. nAChR-coatedJher.7

were pretreated with

PBS + 0. I mg ml-’ BSA followed by the addition of FITCa-NTat

the indicated concentrations. Symbols and bars are

means of triplicate measurements k SEA4 Each data point was determined from a separatejber.

Slopes of the linear

portions of each curve varied by 1e.s.~ than 3%.

a-CNTX (Table 3). Ki values for carb were similar (i.e. 4.5,13.0 or 8.9 mM) irrespective of the labeled toxin used as an indicator of receptor occupancy. Ki values for d-TC were also close (88,52 PM) as measured using FITC-a-BGT and FITC-a-NT; however, the value was only 0.65 FM when using FITC-a-CNTX as the probe. The binding of FITC-a-CNTX to the nAChR immobilized on the optic fiber, was reversible (Fig. 4). Equilibrium binding was established in less than 5 min, and a decremental decrease in the optical signal, by increasing concentrations of &TC. indicated displacement of FITC-a-CNTX by d-TC in a dose-dependent manner. Furthermore, after removal of d-TC and FITC-a-CNTX by an extensive buffer wash (i.e. 10 min), between 70% and 100% of the signal could be recovered. DISCUSSION Immobilization of the nAChR protein on the quartz fiber exposed to solutions of ligands in a 512

(1991) 507-516

flow cell configuration might be expected to affect ligand-receptor binding kinetics. In binding assays, usually the receptors are immersed in a large excess of ligand in a constant volume (O’Brien, 1986). Nevertheless, the relative association and dissociation rates of the three FITClabeled toxins appear to have similar rank order as those reported for toxin interactions with nAChR in its membrane-bound form or in detergent extracts of electric organs (Table 1) (Colquhoun & Rang, 1976; Schmidt & Raftery, 1974). The association rate constant value for the binding of a-BGT to membrane-bound receptors ranges from 2 X IO6MM’min-’ for rat muscle (Colquhoun & Rang, 1976) to 2 X 10’ M-’ min- ’ for Torpedo electroplax (Franklin & Potter, 1972). The observed association rate of a-BGT to nAChR immobilized on the fiber (Table 1) is in the same range. The value reported for a-BGT dissociation from membrane-bound nAChR in rat muscle homogenates was 1.7 X 10e4 min-’ (Colquhoun & Rang, 1976) which is a slower dissociation rate than that of FITC-a-BGT from the receptor-coated fiber (Table 1). The association rate constant reported for aNT binding to membrane-bound nAChR of eel electric organ ranged from 7-6 X IO6 to 2.5 X 10’ M-’ min-’ whereas the dissociation rate constants ranged from 3-2 X 10m3 to 3.9 X 10m3 min-’ (Weber & Changeux. 1974a; Maelicke etal., 1977). Thus. while the association rate of FITCa-NT to nAChR immobilized on the optic fiber is very similar (Table l), the dissociation rate constant (k-‘) is about 14 times larger. These differences in dissociation rate may be due to the source of the receptor (eel versus Torpedo) or the environment (i.e. membrane-bound versus pure protein immobilized on the fiber). Since the kinetic behavior of FITC-labeled a-NT is identical to that of ‘251-a-NT (Johnson & Taylor, 1982). the differences we observed in dissociation rates were probably not due to fluorescent labeling of the probe. The equilibrium dissociation constant Kd for FITC-a-CNTX (5.7 X IO-’ M, Table 1) is in excellent agreement with Kd values determined from bioassays of 2 X lo-’ M (Hashimoto et al., 1985). There are no rate constants reported for a-CNTX, possibly because of the difficulty in measuring its reversible binding to nAChR. It suggests that immobilization of receptors on the optic fiber may be a useful strategy to study kinetics of fluorescent ligand binding to receptors.

Biosensom & Bioektronics

6 (1991) 507-5 16

Pharmacological specific@ of an optical sensor

TABLE 2 Effect of drugs on the specific binding of FIT&z-NT

Muscarinic cholinergic (+) cis-Dioxolane Atropine

fibers

Percentage of inhibtion of initial rate offluorescence changg

DrW

Nicotinic cholinergic Acetykholine Nicotine d-TC

to nAChR-coated

ligands 30 f 3 25 f 8 58 + 7 ligands 20+ 10 0

GABAergic ligands GABA Picrotoxin

0 0

Glutamatergic ligands L-Glutamate 2-Amino-5-phosphonovalerate

0 0

Serotonergic ligands 5-HT ICs 205-930 Adrenergic Norepinephrine Propanalol Dopaminergic Dopamine Haloperidol

23 + 8 0 0 0 0 56 f 8

“The nAChR-coated fibers were pretreated for 5 min with the indicated drugs (1 mM)in PBS + BSA (0. I mg ml-‘) followed by addition of 5 nM FITC-a-NT. hInhibition values are the means f. standard error of the mean of three fibers calculated relative to the rate in the absence of any drug. A value of zero represents nonsignificant difference from the mean.

In summary, the rate constants of association and dissociation ofthree receptor ligands, determined by this optic fiber biosensor, are consistent with the conclusion that this nAChR optical biosensor exhibits a pharmacological profile similar to that observed for the membrane bound receptor. The relative molar fluorescence yields of the three toxins are different. This is believed to be due to the number of available amino groups, which act as conjugation sites for fluorescein. Structural formulae of the three peptide toxins reveal nine amino groups in a-BGT, eight in a-NT and one in a-CNTX. No stoichiometric analysis was performed to determine the number of fluorescein residues per toxin molecule. However, a similar synthetic protocol has been reported for FITC-cz-NT and resulted in one fluorescein residue per a-NT molecule (Johnson

&Taylor, 1982). Although differences in quantum yield may partially explain why picomolar concentrations of a-neurotoxins (i.e. FITC-aBGT, FITC-a-NT) produced optical signals similar to micromolar concentrations of aCNTX, differences in affinity for the receptor (Table 1) are believed to be the more important factor. The Kd values calculated for the three toxins (Table 1) are generally in good agreement with those reported using radioactive ligand binding (Franklin & Potter, 1972; Schmidt & Rafter-y. 1974; Colquhoun & Rang, 1976). Specific binding of FIT&z-NT to the immobilized nAChR protein was dose dependent (Fig. 3). Both the initial rates and the steady state level of the fluorescent signal exhibited dose dependency, as was observed previously for initial rates of FITC-cr-BGT binding for the 513

Biosensors & Bioelectronics 6 (I 991) 507-5 16

Kim R. Rogem et al. TABLE 3 Effect of drugs on the binding of fluorescent-labeled toxin indicators Fluorescent indicator”

Drug

Klb

FIT&z-BGT

a-BGT d-TC Carb

6.0 nM 45

FITC-a-NT

a-NT d-TC Carb

10 nM 52)JM 13 rnM

FITC-a-CNTX

a-CNTX

1+5pM

d-TC Carb

0.65 PM 8.9 mM

8.8~~ rnM

“Assays were performed as described in the Materials and Methods section. FITC conjugates were used at concentrations of 5 nM for FIT&r-BGT and FITC-aNT and 1 PM for FIT&a-CNTX bICM values were determined from Scatchard plots, of the displacement curves. Then K, values were calculated from the equation K, = IC&(l + [I]/KJ.

sensor (Rogers et al., 1989). Displacement of the FITC-a-NT or a-BGT by carb and d-TC results in dose-response functions that have the expected rank order(i.e. (I-TC is more potent than carb), however, carb had an unexpectedly high ICw value. This suggests that the immobilized receptor may have a lower affinity for this drug. The Ki values for carb, calculated from the displacement curves using FITC-a-BGT or FITCa-NT, are one to two orders of magnitude lower than their Ki values obtained from radioactive optical

ligand binding assa! (Table 3) (Weber & Changeux, 19743; Ma, ,-ke etal., 1977). In contrast to car. , hich shows a probeindependent aftir.ny to the receptor, when the reversible FITC-a-CNTX is used to generate the optical signal, the receptor has much higher affinity for d-TC (Table 3). Thus, one may conclude that more reversible probes will be advantageous for use in this system. The sensitivity of the nAChR optical sensor to nicotinic cholinergic drugs and its insensitivity to noncholinergic drugs (Table 3) is further evidence of the pharmacological specificity of the biosensor. Using the quasi-irreversible a-NT as probe requires use of fairly high drug concentrations to displace FTTCa-NT. When the more reversible a-CNTX, is used then much lower drug concentrations are needed. The nAChR protein is composed of five peptides (a&4) and carries a binding site for ACh on each of the a subunits (Hucho, 1986). Carb, d-TC, the snake neurotoxins and a-CNTX bind to these ACh recognition sites. The nAChR also carries sites associated with its ionic channel and these have high affinity for noncompetitive blockers of the receptor, (e.g. local anesthetics (Heidman & Changeux. 1978), histrionicotoxin (Eldefrawi et al., 1980~). phencyclidine (Eldefrawi et al., 1980b) and numerous other drugs (Eldefrawi et al., 1982)). The observed inhibition of FTTC-czNT binding by haloperidol and other nonnAChR drugs may be a result of their binding to the noncompetitive high-affinity site. Inhibition of FITC-a-NT by haloperidol is particularly significant. Haloperidol does not

d Tubocurarine

Buffer Wash

FITC-a-CNTX II l,MI

FITC-a-CNTX (1 UM)

Fig. 4. Inhibition of reversible binding of FLY’-a-CNTX to nAChR immobilized on a quartzfiher by addition of d-TC. The n.AChR-coatedJber waspreincubated with PBS + 0. I mg ml-’ BSA prior to addition of various concentrations of d-TC. After equilibrium was established upon addition of 100 p Md-TC. the receptor-coatedjiber could be reused afrer a IO min wash with PBS. 514

Biosensors & Bioelectronics 6 (1991)

507-516

bind to the ACh recognition site but it binds to the high-affinity noncompetitive binding site, like other antipsychotic drugs (Eldefrawi et al., 1982). This result suggests that fluorescein-labeled channel drugs may also be useful probes. Binding of the channel probe [3H]perhydrohistrionicotoxin to membrane-bound Torpedo is conformation dependent (Eldefrawi et al., 1980a). Receptor agonists increase the association rate of the channel probes by two to three orders of magnitude whereas antagonists do not. Thus, a fluorescent channel probe with binding properties such as perhydmhistrionicotoxin may be more informative than probes which bind to the ACh binding site. TJse of channel probes would make it possible to measure a much faster association rate in the presence of ACh especially if the probe is a selective open channel blocker.

Pharmacological

spectficity of an optical sensor

toxin to acetylcholine Pharmacol.,

receptors in rat muscle. Mol.

12, 519-35.

Cuatrecasas, P. & Hollenberg, M. D. (1975). Binding of insulin and other hormones to non-receptor materials - saturability, specificity and apparent negative cooperativity. Biochem. Biophys. Res. Commun..

62,31-41.

Davis, K. A. & Leary, T. R. (1989). Continuous liquidphase piezoelectric biosensor for kinetic immunoassays. Anal. Chem., 61, 1227-30. Eldefrawi, M. E., Aronstam, R. S., Bakry, N. M.. Eldefrawi. A. T. & Albuquerque. E. X. (1980a). Activation. inactivation and desensitization of the acetylcholine receptor channel complex detected by binding of perhydrohistrionicotoxin. Proc. Nat1 Acad. Sci. USA, 77, 2309- 13.

Eldefrawi. M. E. & Eldefrawi, A. T.. Aronstam, R. S., Maleque, M. A., Wamick. J. E. & Albuquerque, E. X. (1980h). [3H]Phencyclidine: A probe for the ionic channel of the nicotinic receptor. Proc. Nat1 Acad. Sci. USA, 77, 7458-62.

ACKNOWLEDGEMENTS The authors acknowledge the generous support of this research by ORD, Inc. (North Salem, NH) which supplied the instrumentation, and partial financial support (Grant DAAA-15-88-C-0026) and valuable suggestions throughout the course of the study. They also thank Dr A. Galdes of BOC Group Inc. for donating a-Conotoxin GI. Dr A. T. Eldefrawi for help and constructive criticism and MS Sharon Mills for word processing. This work was partially financed by a grant from DOD (DA&I-15-89-RO031) to MEE.

Eldefrawi. A. T.. Miller. E. R.. Murphy. D. L. & Eldefrawi, M. E. (1982). (ZHjPhencyclidine interactions with the nicotinic acetylcholine receptor channel and its inhibition by psychotropic, antipsychotic. opiate, antidepressant, antibiotic, antiviral and antiarrhythmic drugs. Mol. Pharmacol.. 22, 72-81.

Eldefrawi. M. E.. Sherby. S. M.. Andreou. A. G.. Mansour. N. A.. Annau. Z., Blum, N. A. & Valdes. J. J. (1988). Acetylcholine receptor-based biosensor. Anal. Len.. 21, 1665-80.

Franklin. G. I. & Potter. L. T. (1972). Studies of the binding of a-bungarotoxin to membrane-bound and detergent-dispersed acetylcholine receptors from Torpedo electric tissue. FEBS Len.. 28, IOI6.

Gernet. S.. Koudelka, M. & De Rooij, N. F. (1989). A planar glucose enzyme electrode. Sens. Actuators, REFERENCES

17, 537-40.

Albuquerque, E. X.. Eldefrawi, A. T. & Eldefrawi. M. E. (1979). The use of snake toxins for the study of the acetylcholine receptor and its ion-conductance modulator. In Snake Venoms. ed. C.-Y. Lee. Springer-Verlag. Berlin, pp. 377-402. Andrade, J. D.. Vanwagenen. R. A., Gregonis, D. E.. Newby, K. & Lin, J. N. (1985). Remote fiber-optic biosensors based on evenescent-excited fluoroimmunoassay: Concept and progress. IEEE Trans. Electron Devices, ED-32, I 175-9. Birdsall, N. J. M. & Hulme, E. C. (1976). Biochemical studies on muscarinic acetylcholine receptors. J. Neurochem..

27, 7- 16.

Buch, R. M. & Rechnitz. G. A. (1989). Neuronal biosensors. Anal. Chem.. 61, 533A-42. Colquhoun, D. & Rang. H. P. (1976). Effects of inhibitors on the binding of iodinated a-bungaro-

Glass, T. R., Lackie, S. & Hirschfeld, T. (1987). Effect of numerical aperture on signal level in cylindrical waveguide evanescent fluorosensors. Appl. Optics, 26, 2181-7.

Goldstein,

A.. Aronow.

Principles

of Drug

L. & Kalman. S. M. (1969). Hoeher, New York,

Action.

pp. 136-45. Gotoh. M.. Tamiya. E., Momoi. M., Kagawa. Y. & Karube. 1. (1987). Acetylcholine sensor based on ion sensitive field effect transistor and acetylcholine receptor. Anal. Lett.. 20 857-70. Gotoh, M.,Tamiya. E., Seki, A., Shimizu. 1.& Karube. I. (1989a). Glucose sensor based on an amorphoussilicon ISFET. Anal. Lett.. 22, 309-22. Gotoh, M.. Suzuki, M.. Kubo, I., Tamiya, E. & Karube. 1. (19896). Immuno-FET sensor. J. Mol. Catal.. 53, 285-92.

Guilbault.

G. G. & Luongm. J. H. (1989). Acetylcholine 515

Kim R. Rogers et al.

sensor based on ion sensitive field effect transistor and acetylcholine receptor. Se/. El. Rev.. f 1,3-16. Hashimoto, K., Uchida, S., Yoshida, H., Nishiuchi, Y.. Sakakibara, S. & Yukari, K. (1985). Structureactivity relations ofconotoxins at the neuromuscular junction. Eur. J. Pharmacol., 118, 351-4. Heidmann. T. & Changeux, J.-P. (1978). Structural and functional properties of the acetylcholine receptor protein in its purified and membrane-bound states. Annu. Rev. Biochem.. 47, 317-57. Hirschfeld. T. B. & Block, M. J. (1984). Fluorescent immunoassay employing optical fiber in capillary tube. US Patent 4,447.546. Hsiue. G. H., Liu. C. H. & Wang, C. C. (1989). Preparation of glucose oxidase membrane for the application of glucose sensor by plasma activation. J. Appl. Polym. Sci.. 38, 1591-605.

Hucho, F. (1986). The nicotinic acetylcholine receptor and its ion channel. Eur. .I Biochem.. 158, 21 I26.

Johnson. D. A. & Taylor, P. (1982). Site-specific fluorescein-labeled cobra a-toxin. J. Biol. Chem.. 257, 5632-6.

Maelicke, A., Fulpius. B. W.. Klett, R. P. & Reich, E. (1977). Acetylcholine receptor. Responses to drug binding. J. Biol, Chem.. 252, 481 l-30. O’Brien, R. A. (1986). Receptor Binding in Drug Research. Dekker. New York. Rechnitz. G. A. (1987). Biosensors: An 0verview.J. Clin. Lab. Annal.. 1, 308-12.

Rishpon. J. & Rosen. 1. (1989). The development of an immunosensor for the electrochemical determination of the isoenzyme LDHS. Biosensors, 4, 61-74.

516

Biosensors & Bioelectronics 6 (1991) 507-516

Rogers, K. R.. Valdes, J. J. & Eldefrawi, M. E. (1989). Acetylcholine receptor fiber-optic evanescent fluorosensor. Anal. Biochem.. 182, 353-9. Scheller, F.. Schubert, F.. Pfeiffer, D.. Hintsche, R., Dramsfeld, I., Renneberg, R., Wollenberger, U., Riedel. K. & Pavlova. M. (1989). Research and development of biosensors. Analyst (London). 114, 653-62.

Schmidt. J. & Raftery, M. A. (1974). The cation sensitivity of the acetylcholine receptor from Torpedo califomica. J. Neurochem.. 28, 617-23.

Sepaniak. M. J.. Tromberg. B. J. & Vo Dinh. T. (1989). Fiber optic affinity sensors in chemical analysis. Prog. Anal. Spectrosc.. 11, 481-509. Taylor. R. F., Marenchic. 1. G. & Cook, E. J. (1988). An acetylcholine receptor-based biosensor for the detection of cholinergic agents. Anal. Chim. Acta, 213, 131-8.

Weber, M. & Changeux, J.-P. (1974a). Binding of Naja ni&icollis [3H] a-toxin to membrane fragments from Electrophoms and Torpedo electric organs. I. Binding of the tritiated a-neurotoxin in the absence of effector. Mol. Pharmacol.. 10, l-14. Weber, M. & Changeux, J.-P. (19746). Binding of Naja nigricollis rH] a-toxin to membrane fragments from Electrophorus and Torpedo electric organs. II. Effect of cholinergic agonists and antagonists on the binding of the tritiated a-neurotoxin. MO/. Pharmacol., 10, 15-34. Wise, D. L. (1989). Applied BiosensoKy. Butterworths. Stoveham. MA. USA. Wu, C. H:.& Narahashi. T. (1988). Mechanism of action of novel marine neurotoxins on ion channels. Annu. Rev. Pharmacol. Toxicoi.. 28, 141-61.