FUNDAMENTAL

AND

APPLIED

TOXICOLOGY

16, 8 lo-820 (199 1)

Acetylcholinesterase Fiber-Optic Biosensor for Detection of Anticholinesterases KIM R. ROGERS,*”

CHENG J. CAO,* JAMES J. VALDES,~ AMIRA T. ELDEFRAWI,* AND MOHYEE E. ELDEFRAWI*‘~

*Department of Pharmacology Baltimore, Maryland

and Experimental Therapeutics, University of Maryland, 21201; and tBiotechnology Division, U.S. Army Research and Engineering Center, Edgewood, Maryland 21010

Received

September

13, 1990; accepted

January

15, 1991

Acetylcholinesterase Fiber-Optic Biosensor for Detection of Anticholinestemses. CAO,

C. J., VALDES,

J. J., ELDEFRAWI,

A. T., AND

ELDEFRAWI,

School of Medicine, Development

M.

E. (1991).

ROGERS,

K. R.,

Fundam.

Appl.

16, 8 10-820. An optical sensor for anticholinesterases (AntiChEs) was constructed by immobilizing fluorescein isothiocyanate (FITC)-tagged eel electric organ acetylcholinesterase (AChE) on quartz fibers and monitoring enzyme activity. The pHdependent fluorescent signal generated by FITC-AChE, present in the evanescent zone on the fiber surface, was quenched by the protons produced during acetylcholine (ACh) hydrolysis. Analysis of the fluorescence response showed Michaelis-Menten kinetics with a Kapp value of 420 pM for ACh hydrolysis. The reversible inhibitor edrophonium (0.1 mM) inhibited AChE and consequently reduced fluorescence quenching. The biosensor response immediately recovered upon its removal. The carbamate neostigmine (0.1 mM) also inhibited the biosensor response but recovery was much slower. In the presence of ACh, the organophosphate (OP) diisopropylfluorophosphate (DFP) at 0.1 mM did not interfere with the ACh-dependent fluorescent signal quenching, but preexposure of the biosensor to DFP in absence of ACh inhibited totally and irreversibly the biosensor response. However, the DFPtreated AChE biosensor recovered fully after a lo-min perfusion with pralidoxime (2-PAM). Echothiophate, a quaternary ammonium OP, inhibited the ACh-induced fluorescence quenching in the presence of ACh and the phosphorylated biosensor was reactivated with 2-PAM. These effectsreflected the mechanism of action of the inhibitors with AChE and the inhibition constants obtained were comparable to those from calorimetric methods. The biosensor detected concentrations of the carbamate insecticides bendiocarb and methomyl and the OPs echothiophate and paraoxon in the nanomolar to micromolar range. Malathion, parathion, and dicrotophos were not detected even at millimolar concentrations; however, longer exposure or prior modification of these compounds (i.e., to malaoxon, paraoxon) may increase the biosensor detection limits. This AChE biosensor is fast, sensitive, reusable, and relatively easyto operate. Since the instrument is portable and can be self-contained, it shows potential adaptability to field use. o 1991 society of Toxicol.

Toxicology

Detection of anticholinesterases (AntiChEs) is of major concern to the agriculture chemical industry, which manufacturs carbamate and ’ Current address: Biotechnology Division, U.S. Army Research Development and Engineering Center, Edgewood, MD 21010. * To whom reprint request should be addressed. 0272-0590/91 $3.00 Copyright 0 1991 by the Society of Toxicology. All rights of reproduction in any form reserved.

organophosphate (OP) insecticides accounting for most insecticides in use today. Detection of AntiChEs is of concern also to the regulatory agencies that monitor insecticidal residues in food products and the environment. The Department of Defense is interested in detection of AntiChEs because of the threat of use of OP nerve agents. Further, physicians are also 810

ANTICHOLINESTERASE

interested in detection of these chemicals in body fluids. AntiChEs inhibit the catalytic activity of both plasma cholinesterase and acetylcholinesterase (AChE) in red blood cells and neural tissue (Silver, 1974). AntiChEs are divided into two classes. (1) Reversible AntiChEs include quaternary ammonium compounds (e.g., edrophonium) and carbamate esters (e.g., the myasthenia drugs pyridostigmine and neostigmine) which are hydrolyzed by cholinesterases (ChEs) resulting in carbamylation of the enzyme. Carbamylated ChEs are inactive, but regenerate by the relatively slow process of spontaneous decarbamylation (O’Brien, 1967; Taylor, 1985). (2) Irreversible AntiChEs consist mainly of OP esters which are also hydrolyzed by ChEs, resulting in phosphorylation of the catalytic site. In this case, however, spontaneous dephosphorylation is an extremely slow reaction (t& > 8 hr) and may never occur if the phosphorylated-ChE complex ages (O’Brien, 1960). Dephosphorylation occurs faster in the presence of oximes. Pralidoxime (2-PAM) is a therapeutic drug that is used to reactivate AChE in OP poisoning. OP AntiChEs include therapeutics (e.g., echothiophate) (Foldes et al., 1966) insecticides (e.g., parathion and malathion) (Taylor, 1985), and chemical warfare agents (e.g., soman and sarin) (Ellin, 1982). Inhibition of ChEs has been used for analytical purposes to assay for AntiChEs, notably OPs and carbamate compounds. Techniques used to measure AChE have included colorimetric (Ellman et al., 196 1; Hammond and Forster, 1989) radiometric (Lewis and Eldefrawi, 1974), potentiometric (Baum and Ward, 197 1; Durand et al., 1984), colorimeteric (van der Schoot and Bergveld, 1987), and amperometeric (Morelis and Coulet, 1990) methods. While the potentiometric methods appear to be best suited for incorporation into biochemical sensors, limitations have been noted in the areas of speed (Baum and Ward, 197 1; Durand et al., 1984) and selectivity (Baum and Ward, 197 1).

FIBER-OPTIC

BIOSENSOR

811

Our interest in optical sensors is spurred by certain advantages they have over potentiometric sensors (e.g., no direct electric connections, no drift problem, and suitability for continuous monitoring). In a previous study, we immobilized nicotinic acetylcholine receptor on quartz fibers to produce a receptor biosensor (Rogers et al., 1989). This receptor biosensor was effective in detecting fluorescein isothiocyanate (FITC)-labeled a-neurotoxins (a-bungarotoxin and a-cobratoxin), as well as receptor agonists and other antagonists (Rogers et al., 1991). The optical signal, however, was sensitive to the pH of the assay medium due to the pH dependence in the quantum yield of fluorescein. These results suggested that fluorescein may be used as a proton (H+) sensor, which would be useful for monitoring AChE activity, since hydrolysis of its substrate ACh produces H+ ion (Silver, 1974). In this communication, we report on a biosensor which uses FITC-tagged AChE immobilized on quartz fiber to monitor AChE activity in the presence and absence of AntiChEs. This fiber-optic biosensor is fast, reusable, and has excellent substrate specificity and high sensitivity to AntiChEs. It can monitor oxime-dependent enzyme regeneration as well as enzyme “aging.” METHODS Preparation ofFITC-labeled AChE. AChE from eel was obtained from Sigma Chemical Co. (St. Louis, MO) and labeled with PITC as described by Rogers et al. (1989). AChE was suspended in 5 ml of 50 mM bicarbonate buffer, pH 9.5, containing 5 mg/ml (NH&SO4 and 2.5 mg FITC on celite. After incubation for 15 min, the celite was sep arated by centrifugation and the supernatant loaded onto a Sephadex G-25 column (35 X 2.5 cm) and developed with 5 mM ammonium acetate, pH 6.0. AChE activity, which eluted in the void volume, was assayed by the method of Ellman et al. ( 196 1) and the protein determined by the method of Lowry et al. (I 95 1). The specific activity was unaffected by labeling. Immobilization of the FITC-labeled enzyme. Quartz fibets were silanized by a method similar to that reported by Weibel et al. (1973). The fibers were cleaned in concentrated HCI/MeOH ( 1: 1) for 30 min followed by several washes with glass-distilled water and then incubated in concentrated sulfuric acid for 30 min. The libets were then

812

ROGERS ET AL.

washed with seven changes of water. boiled for 5 min in water, and dried under a stream of Nz The cleaned fibers were silanized for 3 hr in refluxing toluene containing 2% aminopropyl triethoxysilane followed by several washes in methanol then air-dried and stored until further use. The silanized fibers were incubated for 16 hr at 4°C in 0.5 ml phosphate buffer (10 mM), pH 4.0, containing FITC-AChE (50 rg/ml) and I-cyclohexyl-3-(2-morpholinoethyl) carbodiimide metho-p-toluene sulfonate (I mg/ ml). The fibers were removed and washed, and the activated groups capped by incubation in 100 mM glycine for 15 min at 23°C. The fibers were then washed and stored in Krebs/phosphate buffer (NaH2P04, 0.1 mM; NaCl, 120 mM; KCl, 4.8 mM; CaCIZ, 1.3 mM; MgS04, 1.2 mM; pH 7.0). The fibers could be stored for several days at 4°C without substantial loss in activity. The use of a physiologically balanced buffer solution yielded better results than phosphate buffer alone. Apparatus. All experimental work was carried out using a fluorometer designed and built by Ord, Inc. (North Salem, NH). The fiber-optic evanescent fluorosensor apparatus was similar in configuration to that reported by Glass et al. (1987) (Fig. 1). Components ofthis instrument included a 10-W Welch Allyn quartz halogen lamp, a Hamamatsu S-1087 silicon detector, an Ismatec fixed-speed peristaltic pump, a Pharmacia strip chart recorder, and bandpass filters and lenses as previously reported. The quartz fibers, 1 mm in diameter with polished ends, were obtained from Ord, Inc.

The fiber-optic evanescent fluorosensor made use of the evanescent wave effect by exciting a fluorophore just outside the waveguide boundary (excitation wavelength = 485/20 nm). A portion of the resultant fluorophore emission then became trapped in the waveguide and was transmitted back up the fiber. This was detected after transmission through 5 IO-nm LP and 530/30-nm filters. The flow cell allowed the center 47 mm of a 60-mm-long fiber to be immersed in 46 ~1 which was exchanged every 14 sec. Fluorescence measurements. After immobilization of the FITC-AChE, the fibers were placed in the flow cell and perfused with Krebs-phosphate buffer until a stable baseline fluorescence was reached (usually 2-10 min). The dependence on pH of the fluorescence yield of the FITCAChE was determined using both the optical biosensor and a Gilford Fluoro IV spectrofluorometer. For biosensor measurements, the fluorescence response, using FITCAChE-coated fibers, was determined as a function of the pH of the perfusion buffer (5 mM bicarbonate; 5 mM Hepes; 5 tnM citrate; in Krebs-phosphate). Fluorescence measurements with the spectrofluorometer (Fluoro IV) were made using FITC-AChE (10 &ml) in the same buffer. Least-squares analysis of double reciprocal plots were performed using Cricket Graph on a Macintosh II computer. For inhibition studies, fibers were preincubated for 10 min by perfusion with various compounds in 0.1 mM phosphate-buffered Krebs solution in the presence or ab-

DETECTOR

f s8.5 SHlJlTER

530630

WELCH ALLYN

510 LP

-FfBERl 1lfHllCMAMf3EA

FLOW CELL

Rc. 1. Schematic presentation of the optical system used to measure fluorescein. Fiber inset illustrates biochemical model for AChE biosensor.

ANTICHOLINESTERASE sence of acetylcholine (ACh) (1 mM) as specified under Results. ACh was then perfused through the flow cell in the presence or absence of inhibitors. The biosensor assay was performed by interrupting the flow and measuring the pH-induced fluorescence change for a 2-min period. Soluble AChE activity was determined by the method of Ellman et al. (196 I). The enzyme was incubated for 10 min in Krebs-phosphate buffer in the presence of various AntiChEs over the concentration range 1 nM to I mM. Activity was then determined for a 2-min assayperiod in the absence of inhibitor. Assayswere run in triplicate and IC50 values determined graphically. Chemicals. Eel AChE, fluorescein isothiocyanate, ACh, butyrylcholine, and other ChE substrates as well as the antiChEs edrophonium and neostigmine and the reactivator 2-PAM were all purchased from Sigma Chemical. The OP compound echothiophate was a gift from Ayerst Laboratories, Inc. (NY), and the insecticides (bendiocarb, methomyl, parathion, malathion, paraoxon, and dicrotophos) were purchased from Chem Service, Inc. (Westchester. PA).

FIBER-OPTIC

BIOSENSOR

813

PH

FIG. 2. Effect of pH on the fluorescence yield of FITCAChE in solution, measured in the Gilford Fluoro IV spectrofluorometer (0) or immobilized on the optic fiber and measured in the Ord, Inc., spectrofluorometer (m). Symbols represent means and bars represent standard error of the mean of three measurements.

RESULTS The fluorescence yield of FITC conjugated to the AChE was dependent on the pH of the buffer (Fig. 2). For the enzyme in solution, complete quenching of the fluorescence was observed at the lowest pH values. However, when immobilized on the quartz fiber, the fluorescence did not drop below 20% of the maximum even at a buffer pH of 3. Initial experiments indicated that the fluorescent signal generated by FITC-AChE immobilized on the fiber in the evanescent zone was extremely stable and showed little drift for several hours. Addition of the substrate ACh to the phosphate-buffered Krebs perfusion medium resulted in a readjustment of the steady-state fluorescence. These changes were complicated by the direct effect of ACh or choline on the fluorescence yield of the FITC-tagged enzyme. The enzyme activity, however, could be assayed by interrupting the flow of the perfusate and measuring the percentage decrease in the baseline fluorescence during a 2-min period. The presence of ACh in the medium at the start of the assay circumvented these problems. The magnitude of signal reduction was dependent on buffer capacity. The change in fluorescence after 2 min varied linearly with the

log of the phosphate concentration between 0.0 1 and 10 mM. A concentration of 0.1 mM was used to maximize the signal change while stabilizing the initial assay pH. Interruption of the buffer flow, by turning the pump off, allowed the local pH in the vicinity of the FITC-labeled enzyme to drop resulting in fluorescence quenching (Fig. 3B). Resuming the buffer flow allowed the equilibrium to be reestablished. The biosenor response was dependent on substrate (Fig. 3A) and the assay was very stable and could be repeated numerous times on the same fiber without loss in enzyme activity (Fig. 3C). Most likely, several factors influenced the rate of fluorescence decrease upon interruption of the buffer flow. These include the decrease in substrate and increase in product in the local environment of the enzyme, the effect of local pH changes on the turnover number, and the nonlinear relationship between proton concentration and quantum yield of FITC (see Fig. 2). Nevertheless, the observed rate of fluorescence quench was dependent on the initial ACh concentration and yielded a rectangular hyperbola indicative of Michaelis-Menten kinetics (Fig. 4).

814

ROGERS ET AL.

B

A

C Krebs,ACh

K&S

Krebs,ACh

FIG. 3. The change in fluorescence was a result of AChE activity. (A) Steady-state fluorescence in the absence of ACh was unaffected by interruption in the buffer flow. (B) In the presence of 1 mM ACh, fluorescence was quenched when the pump was turned off and protons accumulated. The baseline fluorescence was quickly reestablished when the pump was turned on and the excessprotons were removed by the perfusing substrate solution. (C) Enzyme activity was measured by the amplitude of signal quenching after 2 min. This response was stable after 2 hr.

Kinetic analysis of ACh hydrolysis by the immobilized FITC-AChE, measured using the optic biosensor, yielded an apparent K, (i.e., I&,) value of 0.42 mrvr and a V,,, of 400 mV/ 2 min assay. By comparison, the immobilized and soluble FITC-AChE yielded respective K,,, values of 80 and 50 pM for hydrolysis of ace-

tylthiocholine measured by the method of Ellman et al. (196 1). The specific activity measured for the soluble FITC-AChE was 680 pm01 min-’ mg- ’ for the hydrolysis of acetylthiocholine. Using the assumption that the specific activity of FITC-AChE did not change TABLE 1 EFFECT OF AChE SUBSTRATES OF THE AChE FIBER-OPTIC

300.

ON THE RESFQNSE BIOSENSOR

Relative rate of hydrolysis z.E >e

200 .

0.010

Substrate

0.008 e - 0.006

E iii

100

i

0.004

0.002

:rls

012

34

56

1 IlSl

0

0

1

SUBSTRATE

2

3

CONC

4

(mM)

4. Activity of the immobilized FITC-AChE measured at different substrate (ACh) concentrations. Symbols represent means and bars represent standard errors of the mean of triplicate measurements on each of three fibers (N = 9). The inset is a double reciprocal plot of the data used to calculate K,,, and V,,,, values. FIG.

Acetylcholine Acetyl-&methylcholine Benzyl acetate Butyrylcholine Benzoylcholine n-Butyl acetate Carbamvlcholine

Fiber-optic sensor” loo+ 63f 22f 21+ Cl 70% of the initial control activity was observed (Fig. 7D). Thus, the biosensor can be regenerated following irreversible inactivation by the OP AntiChE DFP. Failure of DFP to produce significant inhibition of the immobilized AChE in the presence of ACh suggested that ACh was protecting the esteratic site from phosphorylation. DFP, as well as most insecticidal or nerve agent OPs, lacks the charged ammonium group that helps compounds like ACh, edrophonium, and neostig-

C ACh

4min

FIG. 5. Reversible inhibition of the fluorescent signal generated by the biosensor in presence of 1 mM ACh (A), after 2 min perfusion with 0.1 mM edrophonium + 1 mM ACh (B). After removal of edrophonium and perfusion with 1 mM ACh again (C) the signal was restored. Arrows indicate times when the pump was turned off. The pump was turned on again after 2 min in each case. Three measurements 2 min apart were recorded for each condition.

816

ROGERS ET AL.

A

B ACh

I

Neostigmine.ACh

C ACh

4min

FIG. 6. Inhibition of the AChE biosensor by neostigmine. (A) The biosensor response to acetylcholine (1 prior to antiChE treatment. The immobilized enzyme was then perfused for 10 min with neostigmine (0. I mM) in the presence of acetylchohne (1 mM) and assayed for activity by interrupting the perfusate flow (B). Subsequently, neostigmine was removed from the perfusate and the fluorescence response assayed at 4-min intervals (C). Arrows indicate interruption in the flow of the perfusate. After 2 min the flow was resumed. mM)

mine bind to the active center via the anionic site. Since echothiophate, a therapeutic OP AntiChE, has a charged choline side chain, it should be able to inhibit the enzyme in presence of ACh. This was verified experimentally (Fig. 8B). Furthermore, the inhibited enzyme was totally reactivated by 2-PAM perfusion for 10 min (Fig. SC). Ability of this AChE biosensor to detect insecticidal AntiChE is shown in Fig. 9. The biosensor was treated with aqueous solutions of several carbamate and OP insecticides in the absence of ACh prior to measurement of enzyme activity. The biosensor can detect submicromolar concentrations of an AntiChElike paraoxon, whereas its inactive parent insecticidal compound parathion was hardly detected even at millimolar concentrations. The two carbamate insecticides bendiocarb and methomyl were both detected at micromolar concentrations (i.e.,