Proc. Natl. Acad. Sci. USA Vol. 89, pp. 4168-4172, May 1992 Biochemistry Photoaffinity labeling of avermectin binding sites from elegans and SUSAN P. ROHRER*t, PETER T. MEINKEt, EDWARD C. HAYES*, HELMUT MROZIKT, AND JAMES M. SCHAEFFER* Departments of *Biochemical Parasitology and tBasic Medicinal Chemistry, Merck Sharp and Dohme Research Laboratories, P.O. Box 2000, Rahway, NJ 07065 Communicated by Edward M. Scolnick, February 19, 1992

ABSTRACT An azido-avermectin analog [4"a-(4-azidosal- A icylamido-e-caproylamido-fi-alanylamido)-4"-deoxyaver- mectin B18; azido-AVMJ was synthesized and used to photo- affinity label avermectin binding sites present in the mem- branes of and . Azido-AVM was biologically active and behaved like a com- petitive inhibitor of [3lHlivermectin binding to C. elegans membranes (K; = 0.2 nM). Radiolabeled azido-AVM bound specifically and with high affinity to C. elegans membranes (Kd = 0.14 nM) and, upon photoactivation, became covalently linked to three C. elegans polypeptides of 53, 47, and 8 kDa. Photoaffinity labeling of a membrane preparation from D. melanogaster heads resulted in labeling of a single major R = HO = Avermectin B18 polypeptide of :47 kDa. The that were covalently OH O HH tagged in these experiments are believed to be associated with R = RI N N avermectin-sensitive chloride channels present in the neuro- Nl' ~ f-~ muscular systems of C. elegans and D. melanogaster. Azido- AVM did not bind to rat membranes and therefore was RI = H = Azldo-AVM selective for the and receptors. RI = 1251 = 1251IAzido-AVM The avermectins are a ofmacrocyclic lactones isolated as natural fermentation products from Streptomyces avermi- B tilis (1, 2). (22,23-dihydroavermectin BIa) is a semisynthetic avermectin analog with unprecedented effi- cacy and breadth of spectrum against nematode and arthro- pod parasites. Since its introduction in 1981, it has had a tremendous impact on veterinary health (3-5) and more recently on health (6, 7). It has been approved for prophylactic use against Dirofilaria immitis, the causative agent ofheartworm disease in dogs, and is widely used for the treatment of livestock against a variety of intestinal nema- todes. It is currently the drug of choice for the control and prevention of human onchocerciasis, commonly known as river blindness, a debilitating tropical disease that affects an estimated 18 million people in Africa, Latin America, and the FIG. 1. Avermectin analogs used in this study. (A) Azido-AVM Middle East. Because the avermectins are without toxic side [4"a-(4-azidosalicylamido-E-caproylamido-p-alanylamido)-4"- effects and are efficacious miticidal and insecticidal com- deoxyavermectin Bia)] and 125I-azido-AVM. (B) Octahydroavermec- pounds (8, 9), they also have been developed for use in crop tin (3,4,8,9,10,11,22,23-octahydroavermectin Bia). protection programs. relatively low mammalian toxicity of this family of com- Although the mode of action of avermectins in target pounds (11). Electrophysiological and biochemical studies is not completely understood, they cause an increase support a role for the avermectins in the modulation of in membrane permeability to chloride ions that results in paralysis (10). Specific, high-affinity avermectin binding sites y-aminobutyrate-gated chloride channels in neu- have been identified and characterized in the free-living ronal tissues (13-16). In contrast, the primary site of action nematode, Caenorhabditis elegans (11, 12). A clear correla- of the avermectins in and appears to be a tion exists between the binding affinities for C. elegans non-y-aminobutyrate-gated chloride channel (17-19). membranes and in vivo efficacy of a series of avermectin A detailed understanding of the mechanism of action of analogs, indicating that the binding site is physiologically ivermectin in nematodes and insects could be obtained important. Specific avermectin binding sites also have been through biochemical isolation and characterization of the identified in mammalian brain tissue; however, the affinity is proteins possessing the drug binding sites. Purification of the lower by a factor of - 100 in brain, which may account for the biologically active receptor is not trivial, due to the low abundance of the receptor in tissues and the of ivermectin to the C. The publication costs of this article were defrayed in part by page charge nearly irreversible binding elegans payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this . tTo whom reprint requests should be addressed. 4168 Downloaded by guest on September 24, 2021 Biochemistry: Rohrer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 4169 receptor. Covalent modification of the receptor proteins with (Pierce) supplied as a 10% (vol/vol) solution was added to a photoactive analog of the drug would facilitate purification give a final detergent concentration of 0.5%. This mixture of these proteins under denaturing conditions. In this report was stirred for 1 hr on ice and then centrifuged for 1 hr in a we describe the identification of specific avermectin-binding Ti 65 rotor (Beckman) at 100,000 x g. The 100,000 x g proteins from a nematode (C. elegans) and an insect (Droso- supernatant was filtered through a 0.22-,tm GV-X Millipore phila melanogaster) by photoaffinity labeling with a radio- filter prior to use in binding assays or labeling experiments. iodinated azidoavermectin analog. [31H]Ivermectin Binding Assay. For measurement of binding of [3H]ivermectin to C. elegans membranes or to a Triton X-100-soluble membrane preparation, =200 pug of MATERIALS AND METHODS protein was used per 1-ml assay. The standard binding assay Chemicals. Ivermectin and 3,4,8,9,10,11,22,23-octahy- (11) was used without modification for both the membrane- droavermectin Bia (Fig. 1) were supplied by Helmut Mrozik bound and detergent-soluble C. elegans tissue preparations. and Tom Shih (Merck Sharp and Dohme Research Labora- '251-Azido-AVM Binding and Crosslinking. Triton- tories). [3H]Ivermectin was labeled at the 22,23-position by solubilized C. elegans membrane proteins were diluted with catalytic hydrogenation with tritium gas to a specific activity 50 mM Hepes (pH 7) to a final protein concentration of 200 of 60 Ci/mmol (1 Ci = 37 GBq). ,ug/ml in 0.1% Triton X-100. This extract was then incubated Azido-AVM was synthesized by acylation of 4"-a-amino- with 0.8 nM 125I-azido-AVM in the dark for 1 hr at 220C. To 5-O-tert-butyldimethylsilyl-4"-deoxyavermectin Bla with flu- remove the unbound avermectin, dextran-coated activated oren-9-ylmethoxycarbonyl (Fmoc)-,3-AlaOH. Both protect- charcoal (final concentration, 0.3%) was added. After 10 min ing groups were removed and the resultant amine was acy- the charcoal was removed by centrifugation at 1000 X g. The lated with Fmoc-E-aminocaproic acid. The Fmoc group was supernatant was filtered through a 0.22-,um GV-X Millipore removed and the amine was acylated with N-hydroxysuc- filter. Triton X-100 was added to a final concentration of 1.0% cinimidyl 4-azidosalicylate. All compounds were purified to and the mixture was maintained in the dark at 22°C for 30 min. homogeneity prior to use. Details of the synthesis of azido- Aliquots (2 ml) of the mixture contained in 20-ml glass AVM, will be described elsewhere (P.T.M., S.P.R., E.C.H., scintillation vials were placed in an ice bath on a rotary J.M.S., M. Fisher, and H.M., unpublished work). . The distance from the surface of the sample to the Azido-AVM was radioiodinated by employing chloramine lamp was 6.5 cm. Crosslinking of 125I-azido-AVM to the C. T (20). Two micrograms of azido-AVM in 1 ,1 of dimethyl elegans receptor was achieved by photolysis for 5 min under sulfoxide was added to 50 pl of freshly prepared chloramine a Spectroline model XX-1SB medium-wave UV lamp (30-W T (3 ug/IA1 in acetone). Carrier-free Na1251 (5 mCi in 15 Al output). Comparable results were obtained with UV expo- of 0.01 M phosphate, pH 8.5) was then added and the sure times of 1-10 min (data not shown). C. elegans proteins reaction was maintained at room temperature for 5 min. The were precipitated in 1.5-ml microcentrifuge tubes by addition acetone was removed under a stream of N2 and replaced with of 4 volumes of methanol, storage at -20°C for 1 hr, and 50 ,ul of methanol. The azido-AVM and the 1251I-azido-AVM centrifugation for 10 min at 14,000 rpm in an Eppendorf were resolved by reverse-phase HPLC. The reaction mixture microcentrifuge. was applied to a C18 column (Vydac, 4.6 mm x 25 cm) and Drosophila Head Membrane Preparation and Affinity La- eluted under isocratic conditions with 84% methanol. Azido- beling. Drosophila melanogaster adults of the Oregon-R AVM and 125I-azido-AVM had retention times of9.4 and 10.5 were collected and frozen at -70°C for at least 1 hr but min, respectively. 125I-azido-AVM was obtained in 25% yield up to 2 weeks. Immediately upon removal from the freezer, and was essentially carrier-free with a specific activity of the , contained either in 100-ml glass culture bottles or 1700 Ci/mmol. 50-ml plastic conical culture tubes, were shaken vigorously in All other compounds were supplied by commercial order to break offtheir heads and then poured through a wire sources. mesh screen in order to separate heads from bodies. The C. elegans Cultures. All studies described in this paper were heads were homogenized in a Dounce homogenizer with 50 performed with tissue extracts prepared from the wild-type mM Hepes (pH 7) and centrifuged for 30 min at 28,000 x g. C. elegans strain N2. Worms were grown in liquid culture in The resulting pellet was resuspended in 50 mM Hepes (pH 7) a 150-liter fermentor with as a source of to give a protein concentration of 0.5 mg/ml. food. Details ofthe conditions used for large-scale cultivation Aliquots (100 ,ul) of the Drosophila membrane suspension and harvesting will be described elsewhere (K. Gbewonyo, were incubated with 0.26 nM 125I-azido-AVM in a final S.P.R., L. Lister, B. Burgess, M. Einstein, D. Cully, and B. volume of 1 ml for 1 hr at 22°C. The membranes were Buckland, unpublished work). centrifuged at 28,000 x g and resuspended in 50 mM Hepes Worms were purified by flotation on 35% (21) and (pH 7) containing 0.5% Triton X-100. Aliquots of the resus- washed extensively with 0.1 M NaCl prior to breakage with pended membranes were distributed into glass scintillation a Manton-Gaulin homogenizer. Sucrose-purified worms (500 vials and UV crosslinking was carried out for 5 min as g) were homogenized in 2 liters of 50 mM Hepes (pH 7) described above. Drosophila proteins were precipitated with containing 1 mM ethylenediaminetetraacetic acid (EDTA), 80% methanol. 0.2 mM phenylmethanesulfonyl fluoride, leupeptin at 0.5 ,g/ml, and pepstatin at 0.7 ,tg/ml. The homogenate was centrifuged at 1000 x g and the pellet was discarded. The RESULTS supernatant was centrifuged at 28,000 x g and the resultant Inhibition of [3H]Ivermectin Binding by Azido-AVM. To pellet was resuspended at a final protein concentration of 5 confirm that the azido-AVM analog retained biological ac- mg/ml in 50 mM Hepes (pH 7) containing the protease tivity, this compound and ivermectin were evaluated in the C. inhibitors listed above. Membrane suspensions were dia- elegans motility assay. The LD95 values were 10 ng/ml and lyzed extensively against the same buffer prior to freezing at 3 ng/ml, respectively. -700C. The IC50 for ivermectin was 0.2 nM and the IC50 for The C. elegans motility assay used for assessing biological azido-AVM was 0.3 nM (Fig. 2A). The biologically inactive activity of avermectin analogs has been described (11). analog octahydroavermectin had no inhibitory effect on Solubilization of the Avermectin Binding Site. Frozen mem- [3H]ivermectin binding at concentrations up to 100 nM. Fig. branes were thawed and diluted with 50 mM Hepes (pH 7) to 2 B and C shows the result from a competitive binding assay a final protein concentration of 2 mg/ml. Triton X-100 where saturable binding of [3H]ivermectin to a 28,000 x g Downloaded by guest on September 24, 2021 4170 Biochemistry: Rohrer et al. Proc. Natl. Acad. Sci. USA 89 (1992) A 120r B 100 10 - 2 c) C c E o No Inhibit 1.0 0 o 80 8 * 2X10-10lx1O-10 E X- o-* 4x1010 0.8 -cO i-. 6- = 60 _0:3 0 a) 0.6 -0- IVM 4- .0m .0 0 -A- Azido AVM M 40 (n 0.4- -* OctahydroAVM 2 I3 20 0.2 0- lllllllllllllll -7 -5 -3 -1 1 3 5 7 9 11 0.0

-00 -10 -9 -8 -7 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 log10[inhibitor] (M) [3H]IVM, nM Azido-AVM, nM

FIG. 2. Inhibition of[3H]ivermectin ([3H]IVM) binding to C. elegans membranes by azido-AVM. (A) C. elegans 28,000 x g membrane protein (-200 jig in a 1-ml assay volume) was incubated with 1 nM [3H]IVM in the absence or presence of various concentrations of unlabeled IVM, azido-AVM, or octahydroavermectin. (B) C. elegans 28,000 x g membrane protein (-250 ,ug in a 1-ml assay volume) was incubated with various concentrations of [3H]IVM in the absence (o) or presence of three different fixed concentrations [0.1 nM (e), 0.2 nM (U), or 0.4 nM (o)] of unlabeled azido-AVM. (C) Replotting the data shown in B gave a K1 of0.2 nM. Incubations and filter binding assays were performed in the dark.

membrane fraction from C. elegans was measured in the ity-labeled protein appearing at 12 kDa on this gradient gel presence of0.4, 0.2, 0.1, and 0.0 nM azido-AVM. The double was analyzed in an SDS/20% polyacrylamide gel and found reciprocal plot (Fig. 2B) establishes that azido-AVM is a to be closer to 8 kDa (data not shown). competitive inhibitor of ivermectin, with Ki = 0.2 nM (Fig. Fig. 4C shows the autoradiogram from a similar experiment 2C). The experiments described in Fig. 2, as well as the C. in which the ability of ivermectin to block affinity labeling of elegans motility assay, were performed in the dark in order to the receptor by 125I-azido-AVM was compared with the effect avoid photoactivation of the azido-AVM analog. of octahydroavermectin, the inactive analog. While ivermec- 1I25-Azido-AVM Binding. Saturable binding of 1251I-azido- tin competed effectively for the avermectin binding site and AVM to a C. elegans 28,000 X g membrane preparation was prevented labeling by 125I-azido-AVM, octahydroavermectin demonstrated (Fig. 3A). Scatchard analysis of the data gave a was ineffective. This finding is consistent with previously Kd for 125I-azido-AVM of 0.14 nM and a B. of0.38 pmol/mg established octahydroavermectin binding data (11) and fur- (Fig. 3A Inset). These numbers agree closely with findings for ther substantiates the claim that the three C. elegans proteins [3H]ivermectin binding to C. elegans membranes (11). labeled by 125I-azido-AVM were indeed labeled specifically Photoaffinity Labeling of C. elegans Avermectin Binding and represent the high-affinity binding site. Sites. Aliquots of the Triton X-100-solubilized C. elegans Photoaffinity Labeling ofD. melanogaster Avermectin Bind- membrane proteins were assayed at each step of the binding ing Sites. Affinity labeling of Drosophila head membranes and crosslinking protocol in order to ensure that the ratio of was performed with 0.26 nM 1251I-azido-AVM in the absence specific to nonspecific binding of 125I-azido-AVM to C. or presence of 10 nM unlabeled ivermectin. Fig. 5 shows that elegans proteins did not change at any one step in the a single major protein with estimated molecular mass of 47 experiment. Fig. 3B shows the result obtained when samples kDa was labeled in a specific fashion. A minor protein band were incubated with the 125I-azido-AVM in the absence or near the bottom of the gel (15 kDa) was also labeled. On this presence of various concentrations of unlabeled ivermectin autoradiogram a lane containing free 125I-azido-AVM shows and then subjected to UV crosslinking. The profile resembled that the compound runs well below the smaller protein bands the competition curve seen in Fig. 2A, indicating that there that are labeled in both the C. elegans and the Drosophila was no increase in the nonspecific binding of 1251-azido-AVM preparations. to C. elegans proteins as a result of UV crosslinking. Effi- ciency ofcrosslinking was estimated at 23-30% by comparing DISCUSSION the amount of radioactivity present after precipitation of proteins with methanol to the amount of radioactivity bound A biologically active azido-AVM analog has been synthe- by high affinity to the Triton-soluble C. elegans proteins at sized and radiolabeled. In the dark, this molecule acted as a the initial binding step prior to charcoal adsorption and UV competitive inhibitor of ivermectin binding to C. elegans crosslinking. membranes. High-affinity avermectin-binding proteins from Fig. 4A shows a Coomassie-stained gel ofC. elegans Triton C. elegans and D. melanogaster were covalently labeled by X-100-solubilized membrane proteins used in the affinity UV crosslinking with 125I-azido-AVM. In both tissue prep- labeling experiments. Equal loading of protein samples as arations, binding as well as affinity labeling was blocked by demonstrated here occurred in both of the experiments the addition of competing unlabeled ivermectin, whereas a displayed in Fig. 4 B and C. biologically inactive analog, octahydroavermectin, failed to Fig. 4B shows on an autoradiogram the experimental result block the labeling of the C. elegans proteins. The unambig- obtained and described in Fig. 3B. In the absence of unla- uous results obtained with both C. elegans and D. melano- beled, competing ivermectin, three major proteins with es- gaster have demonstrated the presence of discrete and spe- timated molecular masses of approximately 12, 47, and 53 cific high-affinity binding proteins in the membranes of these kDa were labeled. As the amount of competing unlabeled . Azido-AVM was highly selective for the nem- ivermectin added during the incubation with 125I-azido-AVM atode and insect receptors and did not specifically label any was increased (lanes 2-5), the labeling of the bands with mammalian brain membrane proteins. 125I-azido-AVM was diminished. These results demonstrate Three major polypeptides were labeled in the C. elegans that the three bands were labeled specifically and were tissue preparation, whereas only a single major species was indicative of the high-affinity binding site. The smaller affin- labeled in the Drosophila membranes. Whether the three Downloaded by guest on September 24, 2021 Biochemistry: Rohrer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 4171 B A 1 2 3 45 C 1 2 3 4 5 A E M 1 2 3 4 5 0 E Q. kDa ~0a- 97.4 e, -0I 66.2 a-n 12345H

CM 45.0 d-I -0 *0 31.0 b-

21.5 0- 14.4 be 1251-azido-AVM, nM FIG. 4. Photoaffinity labeling of the Triton X-100-soluble form of the C. elegans avermectin receptor with 125I-azido-AVM. Triton B X-100-soluble C. elegans membrane proteins were incubated with 120 0.8 nM 125I-azido AVM in presence of various concentrations of unlabeled ivermectin. (A) Representative Coomassie stained-gel showing the C. elegans proteins after electrophoresis in an SDS/5- -°100 20% polyacrylamide gradient gel. (B) Autoradiogram of the same gel. 0 Two hundred micrograms ofTriton X-100-soluble membrane protein was loaded per lane. C. elegans proteins were labeled with 1251_ ~0I- 80lF D0 0 azido-AVM in the absence of unlabeled ivermectin (lane 1) or in the .060 presence of0.2, 0.8, 2, or 20 nM ivermectin (lanes 2-5, respectively). L- Arrowheads at right in A indicate where the affinity-labeled bands shown in B line up on the Coomassie-stained gel. Lane M, size *0 markers. (C) Autoradiogram from a separate experiment. Conditions -a- IVM for affinity labeling and electrophoresis were the same as in the -_- Octahydro AVM experiment shown in A and B. Proteins were labeled with 125I-azido- AVM in the absence of unlabeled ivermectin (lane 1), in the presence in 20 of 20 or 2 nM ivermectin (lanes 2 and 3), or in the presence of 20 or 2 nM octahydroavermectin (lanes 4 and 5).

-so ''-1 0 -9 -8 -7 However, covalent tagging of the C. elegans avermectin log1o[inhibitor] (M) binding site with 125I-azido-AVM has eliminated the need to purify under nondenaturing conditions and has greatly in- FIG. 3. (A) Specific binding of 125I-azido-AVM to C. elegans creased the probability of purifying the binding-site proteins membranes. C. elegans 28,000 x g membrane protein (250 jig in to homogeneity. 1 ml) was incubated in the dark with various concentrations of The cloning and structure determination of this binding 1251I-azido-AVM. Specific binding was determined by subtracting the site, which presumably is part of an avermectin-sensitive amount of 125I-azido-AVM bound in the presence of 1 gsM ivermectin chloride channel, should prove useful for understanding of from the amount of 125I-azido-AVM bound in the absence of unla- beled ivermectin. (Inset) Scatchard plot of the saturation curve from the mode of action of the avermectins. Information gathered which Bmax and Kd values were obtained. (B) Inhibition of 125I-azido- from such studies may facilitate the discovery or design of AVM binding by ivermectin (IVM). Triton X-100-soluble C. elegans membrane protein (200 jig) was incubated in the dark with 0.8 nM 1 2 3 4 125I-azido-AVM in the presence of various concentrations of unla- beled ivermectin. Specific binding of the 125I-azido-AVM to the detergent-soluble form of the C. elegans receptor was determined by performing the filter (Whatman GF/B) binding assay both prior to UV crosslinking and after UV crosslinking. The result obtained after UV crosslinking is shown. polypeptides from C. elegans represent degradative products of a single binding protein, individual subunits of a hetero- trimeric receptor complex, or tissue-specific forms of high- affinity avermectin-binding proteins could be determined by isolating and sequencing each product. The apparent number of avermectin binding sites present on Drosophila mem- branes was 5-10 times higher than that in the C. elegans tissue preparations and probably reflects the enriched pop- FIG. 5. Photoaffinity labeling of the Drosophila avermectin- ulation of neuronal membranes in the Drosophila head binding protein with 1251I-azido-AVM. A 28,000 x g membrane homogenate. fraction prepared from Drosophila heads was incubated with 0.26 nM Prior to initiating the affinity labeling approach, we di- 1251-azido-AVM in the absence or presence of 10 nM unlabeled ivermectin. Proteins were electrophoresed in an SDS/5-20% poly- rected substantial effort toward purification of the biologi- acrylamide gel and subjected to autoradiographic analysis. Lane 1, cally active C. elegans receptor in its native state. Synthesis free 125I-azido-AVM; lane 2, C. elegans labeling pattern as seen in of the avermectin analogs and affinity resins used in this Fig. 4; lane 3, labeling of the Drosophila head membrane preparation approach will be discussed elsewhere (P.T.M., S.P.R., with 125I-azido-AVM; lane 4, labeling of the Drosophila head mem- E.C.H., J.M.S., M. Fisher, and H.M., unpublished work). brane preparation in the presence of 10 nM unlabeled ivermectin. Downloaded by guest on September 24, 2021 4172 Biochemistry: Rohrer et al. Proc. Nati. Acad. Sci. USA 89 (1992) compounds that target this invertebrate chloride channel and 7. Taylor, H. R., Pacque, M., Munoz, B. & Greene, B. M. (1990) possess properties of even greater efficacy, safety, and Science 250, 116-118. breadth of spectrum than the avermectins. 8. Ostlind, D. A., Cifelli, S. & Lang, R. (1979) Vet. Rec. 105, 168. 9. Putter, I., MacConnell, J. G., Preiser, F. A., Haidri, A. A., We thank Drs. Roy Smith, Mervyn J. Turner, and Michael H. Ristich, S. S. & Dybas, R. A. (1981) Experientia 37, %3-964. Fisher for their interest in and support of this project. We wish to 10. Turner, M. J. & Schaeffer, J. M. (1989) in Ivermectin and acknowledge Monica Einstein and Easter Frazier for technical Abamectin, ed. Campbell, W. C. (Springer, New York), pp. assistance and Drs. Joe Arena and Doris Cully for helpful discussions 73-88. throughout the course ofthis study. We thank Dr. Kodzo Gbewonyo, 11. Schaeffer, J. M. & Haines, H. W. (1989) Biochem. Pharmacol. Leonard Lister, and Bruce Burgess for growing large batches of C. 38, 2329-2338. elegans worms. 12. Cully, D. F. & Paress, P. S. (1991) Mol. Pharmacol. 40, 326-332. 1. Burg, R. W., Miller, B. M., Baker, E. E., Birnbaum, J., Cur- 13. Pong, S. S. & Wang, C. C. (1980) Neuropharmacology 19, rie, J. A., Harman, R., Kong, V. L., Monaghan, R. L., Olson, 311-317. G., Putter, I., Tunac, J. D., Wallick, H., Stapley, E. O., Oiwa, 14. Drexler, G. & Sieghart, W. (1984) Eur. J. Pharmacol. 101, R. & Omura, S. (1979) Antimicrob. Agents Chemother. 15, 201-207. 361-367. 15. Drexler, G. & Sieghart, W. (1984) Eur. J. Pharmacol. 99, 2. Egerton, J. R., Ostlind, D. A., Blair, L. S., Eary, D. H., 269-277. Suhayda, D., Cifelli, S., Riek, R. F. & Campbell, W. C. (1979) 16. Sigel, E. & Baur, R. (1987) Mol. Pharmacol. 32, 749-752. Antimicrob. Agents Chemother. 15, 372-378. K. P. S. & D. F. 3. Chabala, J. C., Mrozik, H., Tolman, R. L., Eskola, P., Lusi, 17. Arena, J. P., Liu, K., Paress, Cully, (1991) A., Peterson, L. H., Woods, M. F., Fisher, M. H., Campbell, Mol. Brain Res., in press. W. C., Egerton, J. R. & Ostlind, D. A. (1980) J. Med. Chem. 18. Zufall, F., Franke, C. & Hatt, H. (1989) J. Exp. Biol. 14, 23, 1134-1136. 191-205. 4. Campbell, W. C. (1985) Parasitol. Today 1, 10-16. 19. Martin, R. J. & Pennington, A. J. (1989) Br. J. Pharmacol. 98, 5. Campbell, W. C. (1989) Ivermectin and Abamectin, ed. Camp- 747-756. bell, W. C. (Springer, New York). 20. Tae, H. J. & Inhae, J. (1982) Anal. Biochem. 121, 286-289. 6. Greene, B. M., Brown, K. R. & Taylor, H. R. (1989) in Iver- 21. Sulston, J. & Hodgkin, J. (1988) in The Nematode Caenorhab- mectin and Abamectin, ed. Campbell, W. C. (Springer, New ditis elegans, ed. Wood, W. B. (Cold Spring Harbor Lab., Cold York), pp. 311-323. Spring Harbor, NY), pp. 602-603. Downloaded by guest on September 24, 2021