Tor&at Vol . 29, No. 10, pp . 1195-1204, 1991 . Printed i° Great Britain .

n Ptm pk

PRODUCTION AND CHARACTERIZATION OF MONOCLONAL ANTIBODIES AGAINST NAJA NAJA ATRA COBROTOXIN BRADLEY G. STILES, 1 BURTON C . LII~GERDING, 2 FRANCIS W . SEXTON I

GuFST2

and

$EIAWN B .

'Department of Toxinology, Pathophysiology Division; and Virology Division ; U.S . Army Modical Research Institute of Infectious Diseases, Frederick, MD 21702-5011, U.S .A . (Received 7

April 1991 ;

accepted 30 May

1991)

and S. B . GUEST . Production and characterization of monoclonal antibodies against Naja raja atra cobrotoxin . Toxicon 29, 1195-1204, 1991 .-Twelve monoclonal antibodies against cobrotoxin from Naja raja atra venom were tested for cross-reactivity with eight different snake toxins, binding to linear epitopes, prevention of cobrotoxin binding to acetylcholine receptor (AchR) in vitro, and protection in mice concomitantly given a lethal dose of cobrotoxin . The antibodies were highly specific, as evidenced by little reactivity with other snake toxins . None of the monoclonal antibodies bound to reduced cobrotoxin or synthesized 8-mer regions spanning the whole molecule, thus suggesting the recognition of conformational epitopes . The in vitro binding of toxin to AchR was competitively inhibited (23-79%) with a 1 .66 :1 mole ratio of antibody : AchR . Preincubation of monoclonal antibody with toxin before adding AchR (3 : 1 mole ratio of AchR : antibody) inhibited the in vitro binding of toxin to AchR by 20-80% . Monoclonal antibodies added after the preincubation of toxin with AchR did not dissociate the toxin-AchR complex. An antibody : toxin mole ratio of 2.5 : 1, with 6 rug of cobrotoxin, delayed the time to death of mice 3.7-23 .8-fold over control mice. The monoclonal antibodies that most effectively prevented in vitro binding of toxin to AchR also provided the longest delay in time to death in mice. B . G . STILES, B . C . LIDGERDING, F . W. SE7CTON

INTRODUCTION

isolated from the venom of Naja raja atra, is a short-chain postsynaptic neurotoxin composed of 62 amino acids and is dependent on four disulfide bonds for biological activity (YANG, 1965, 1974 ; YANG et al., 1969, 1970). The long-chain postsynaptic neurotoxins, such as a-bungarotoxin, differ in structure by having 70-73 amino acids and five disulfide bonds (YANG, 1974). Snake venom postsynaptic neurotoxins have homologous amino acid regions and consequently have similar conformations and binding affinity constants for acetylcholine receptor (AchR) (WESER and CHANGEUX, 1974 ; YANG, 1974; WALKINSHAW et al., 1980; DUFTON and HIRER, 1983 ; Low and CORFIELD, 1986). Various investigators have studied the binding of snake venom postsynaptic neurotoxin regions to AchR . JUILLERAT et al. (1982) showed that a peptide loop region (amino acids 1648) of a Naja raja philippinensis short-chain neurotoxin binds AchR with more affinity COBROTOXTN,

1195

1196

B. G. STILES et at .

than acetylcholine, but less than native toxin . A peptide loop region from a-bungarotoùn (amino acids 26-41) also recognizes AchR with lesser affinity than the native toxin (McDArI>EL et al., 1987). Other binding studies with a-bungarotoùn have used AchR peptides from the a subunit (MuLAC-JERICEVIC and ATASSI, 1987 ; Mur..AC-JERICEVIC et al ., 1988), as well as three different loop regions from a-bungarotoùn which include amino acids 3-16, 261, and 45-59 (RUAN et al., 1990). The highest affinity of an a-bungarotoxin peptide (amino acids 26-41) for AchR peptides was approximately 10-fold less than native toxin for the same receptor peptides . Various immunological studies with snake venom postsynaptic neurotoains have demonstrated neutralizing polyclonal antibodies that cross-react with other snake neurotoxins (MENEZ, 1985). One molecule of cobrotoxin or Naja nigricollis toùn a accommodates three or four antibody molecules, respectively (CHANG and YANG, 1969; YANG et al., 1978 ; MENEZ et al., 1979). Antibody against Laticauda semifasciata venom or purified erabutoùn b cross-reacts with many different short-chain postsynaptic neurotoùns (BoQuET et al ., 1973 ; ABE and TAMnrA, 1979), except cobrotoxin, which shares 58% amino acid homology with erabutoxin b (YANG et al., 1969 ; ABE and TAMIxA, 1979). More recent studies have been done with monoclonal antibodies against snake venom postsynaptic neurotoùns (BOULAIN et al., 1982; DANSE et al., 1986 ; ia:iUANG et al., 1989 ; KASE et al ., 1989 ; PACHNER and RICALTüN, 1989). Three different investigations have resulted in monoclonal antibodies that neutralize a-bungarotoùn in vivo (DANSE et al ., 1986; CwuANC et al., 1989 ; KASE et al ., 1989). These antibodies were developed against formylated or native a-bungarotoùn (DANSE et al ., 1986; CY~uANC et al., 1989 ; KASE et al., 1989). A neutralizing epitope on a-bungarotoùn was identified within amino acids 34 1 (CFILJANG et al., 1989) and 34-38 (KASE et al., 1989). This is also the same region of the toxin that binds AchR peptides with the highest affinity (RUAN et al., 1990). The monoclonal antibody which recognized amino acids 34-38 (SSRGK) of a-bungarotoùn also neutralized other long-chain newotoxins containing an identical sequence, but not closely related neurotoùns differing by only one amino acid (S-35 to I-35) (KASE et al.,

1989). The few monoclonal antibody studies done with snake venom short-chain postsynaptic neurotoxins include those of BOULAIN et al. (1982) and TREMFAU et al. (1986) . Both investigations resulted in different monoclonal antibodies against N. nigricollis toùn a which prevented the in vitro binding of toxin to AchR and neutralized the lethal effects of toxin a in mice . The antibodies differed in epitope recognition and specificity, with one antibody binding to seven different short-chain, but no long-chain, neurotoùns (TREMF.AU et al., 1986). Although there are various reports on the neutralizing capabilities of polyclonal antisera against N. n. atra cobrotoxin (GIANG and YANG, 1969, 1973 ; YANG et al., 1977), there is only one report of a monoclonal antibody against cobrotoxin (YANG, 1985) but no characterization work was done regarding binding specificity and toxin

neutralization . This study evaluates 12 monoclonal antibodies against N. n. atra cobrotoxin and their ability to cross-react with other snake venom toùns, prevent in vitro binding of cobra toxin to AchR, neutralize cobrotoxin in vivo, and finally react with linear epitopes on cobrotoxin . MATERIALS AND METHODS Cett lines The non-immunoglobulin-secreting marine plasmacyloma, S1?2/0-Agl4 cell line (Sttut.~wN er d., 1978) and all

Cobrotoxin Monoclonal Antibodies

1197

developed clones were grown in Optimem medium (Gibco/BRL, Grand Island, NY, U.S .A .) with 7.5% fetal bovine serum, 2 mM glutamine, 50 kg/ml gentamycin (Whiltaker Bioproducts, Walkersvilk, MD, U.S.A .), 2 pg/ml Fungizone (Giboo/BRL), and 15% conditioned medium . Conditioned medium was prepared by filtering Eagle's minimum easeatial medium from 7-14-day-old MRC-5 cells through a 0.22-fan filter (Millipore . Bedford, MA, U.S .A.). MRC-5 cells were obtained from the Salk Institute (Swiftwater, PA, U.S .A .), and maintained is antibiotio-free medium .

Monoclonal antibody production Male BALB/c mice (25 g) were each injected i.p . with 1 Kg of N . n. afro cobrotoxin mixed in 200 pl of RIBi adjuvant (Hamilton, MT, U.S .A .). Mice were injected three times, once every week, and antibody filets were tested by an enzyme-linked immunosorbent assay (ELISA) as described below. Spleens were taken from mice killed three days after a final boost and fused with SP2/0-Ag14 cells at a 1 : 2 ratio using 50% (v/v) polyethYlene glycol 1500 MW (Boehringer-Mannheim, Indianapolis, IN, U.S.A.) as the fusogen (F.~x~nr and Osr©eurrc, 1985). Fused cells were plated in 9lrwell plates and grown in Optimem medium containing hypoxanthine, atainopterin, and thymidine (Boehringer-Mannheim) for 14 days to select for antibody-producing cells. Production of antibody against tobrotoxin by hybridomas was detected with an ELISA as described below. Antibody-producing cells were cloned twice by the limiting dilution method of RF.rvea et al. (1985) without a feeder layer. The sub-isotype of the monoclonal antibodies was determined on cell culture fluids by ELISA with monospectifit antisera (Hyclone, Logen, UT, U.S.A .) . Selected cloned hybridomas were grown in BALB/c strain CJ mice to produce oscine fluid. Monoclonal antibodies were purified from ascàtes fluid by affinity chromatography with protein G-agarose (Pharmacie, Piacataway, NJ, U.S .A.), dialysed against phosphate buffeted saline (PBS; pH 7.4) and stored at -70°C.

Frtzymt-linked imnwnosorbcnt assays (ELISA)

Antibody titers against cobrotoxin in mice were monitored with an ELISA by coating each well on an Immulon II miaotiter plate (Dynatech, Chantilly, VA, U.S.A .) with 100 pl of cobrotoxin (2 pg/ml carbonate buffer, pH 9.6) overnight at 4°C. Unoccupied sites in each well were then blocked with 300 pl of PBS containing 1% gelatin (PBSG) for 30 min at 37°C . Mouse sorom was diluted 1 : 100 in PBS containing 0.1 % Tween 20 and 0.1% gelatin (PHSTG), and added (100ît1/well) for lhr at 37°C . Wells were washed four times with PBS containing 0.1 % Tween 20 (PBST), 100 Ill of PHSTG containing anti-mouse antibody conjugated with alkaline phosphatage (Sigma, St Louis, MO, U.S .A .) was added, and incubated for 1 hr at 37°C . Wells were then washed five times with PHST, 100It1 of pare-nitrophenyl phosphate substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD, U.S.A .) was added, and the absorbance was read at 405 nm after 30 min at room temperature . Detection of hybridomas producing antibody against cobrotoxin was accomplished by initially coating each Immulon II nticrotiter well with 100 pl of a 1 l+g oobrotoxinJml carbonate buffer solution (pH 9.6) overnight at 4°C. After blocking unoccupied sites for 30 thin at 37°C with PHSG, 100It1 of undiluted growth medium ooataining antibodies was added per well and incubated 1 hr at 37°C . Wells were then washed twice and anti-mouse antibody conjugated with alkaline phosphatage in PHSTG was added for 1 hr at 37°C . Wells were finally washed four times, substrate added for 30 min, and the absorbance read at 405 tun. The cross-reactivity of cobrotoxin monoclonal antibodies with purified snake venom toxins (toxins were purchased from Natural Product Sciences, Salt Lake City, UT, U.S .A.) wan tested with crotamine and crotoxin from Crotalus drtrisrus terriftcus, cardiotoxin and cobrotoxin from Naja raja afro, ß-bungarotoxin from Bungarvs multicinctus, and erabutoxin b from Laticauda semijasciata . Purified N. n. oxiana neurotoxins I and II, and B. tratlticinctus a-bungarotoxin (Sigma) were also tested. All toxins were adsorbed onto Immuloa II microtiter wells at 4°C overnight with a 10 pg toxin/ml carbonate buffer solution (pH 9.6 ; 100 Ill/well). Monoclonal antibodies (100 Ill/well of a 10 pg/nil PHSTG solution) were then added and the cross-reactivity was determined by an ELISA as previously described for mouse serum.

Acrtylchollne receptor assays Acetyk holine receptor (AchR) assays were done essentially as previously described (Snt.>:s, 1991). Monoclonal antibodies were diluted in PHSTG and 100 p1 was added to wells containing adsorbed cobroloxin (10014 of a l0ltg toxin/ml carbonate buffer solution; pH 9.6) I hr before, during, or after the addition of AchR (50Itglml PHSTG) to the plate. The receptor bound to toxin was detected with a 1 : l00 dilution of guinea-pig anti-AchR incubated for 1 hr at room temperature. After washing twice with PBST, 100It1 of and-guinea-pig antibody alkaline phosphatage conjugate (Sigma) was added per well for an additional 1 hr at room temperature. The plate was finally washed five rimes, substrate added, and absorbance read at 405 nm.

1198

H. G. STILES et al. TABLE l . REACTIVITY OF MONOCLONAL ANTIBODR~3 (I pg/well) WITH COBROTOXIN (I lig~WCII) OD,~ Antibody

Cobrotoxin

Reduced Cobrotoxin

Control'

2A7 lE8 IOD12 1B93 6A 10 1D8 8H5 13H7 13H8 15E1 9B5 ISF2

0.811 ±0.022 0.905±0.013 1 .0I210 .01 l 1.05710 .040 1 .06610 .116 1 .10510 .038 I .11010 .009 1 .29810 .027 I .40010 .079 1 .707±0 .032 1 .79210 .053 I .90210 .020

0.01710 .010 0.08010 .010 0.00010 .000 0.042 ±0 .012 0.01610 .012 0.07610 .003 0.07210 .001 0.01610 .007 0.0161O .O14 0.02410.015 0.085 f0.028 0.08810.004

0.00310.001 0.00410.002 0 .004±0.001 0.01010.003 0.02310.010 0.03310.005 0.01610.002 0.00910.004 0.02810.011 0.025±0.003 0.03910.013 0.00810.002

Control wells contained blocking solution (1% gelatin), monoclonal antibody, conjugate and substrate . Mouse hyperimmune and preimmune sera (1 :100 of each) control readings for native Cobrotoxin were > 2 .800 and 0.05610.006, respectively . Mouse hyperimmune and preimmune sera (1 : 100 of each) control readings for reduced Cobrotoxin were 0.30610.040 and 0.00710.005, respectively . All antibodies were tested in triplicate and the average and S.D. determined .

Neutralization assay Swiss Webster mice (23 g 1 I .8 g) were each injected i.p . with 300 pl of wbrotoxin diluted in PBS (6 ~g/mouse) and monoclonal antibody (200 pg/ml or 1000 pgJml PBS final dilution) . Toxin and antibody were preincubatod together for 30 min at room temperature before the injections.

TABLE 2. INHIBTTION OF TOXIN~ACIIR COMPLEXING (PREINCUBATION OF TOXIN WTTH ANTIBODY BEFORE AchR) Antibody 1B93 1SE1 2A7 6A10 13H7 13H8 ISF2 1OD 12 9B5 IB92 1D8 BH5 lE8

10 pg/ml (% Comp)" 1 .67110 .066 1 .65210 .048 1 .63210 .060 1 .47410 .056 1 .39510 .013 1 .37710 .112 1 .25110 .050 1 .096 t0.163 1 .07410 .271 0.81810 .152 0.65210 .124 0.45710 .041 0.41510 .096

(20) (21) (22) (29) (33) (34) (40) (47) (48) (61) (71) (78) (80)

100 ng/ml (% Comp) 2.087±0 .188 1.92710 .054 1.87010.030 2.00810.I13 1.91110.139 2.00810.106 I.92110.072 I.83110.134 1.98810.045 1.58810.095 l.79810.066 1.73210.167 1.76510.014

(0) (4) (10) (4) (9) (4) (8) (12) (4) (9) (14) (17) (15)

1 ng/ml (% Comp) 2.3I810 .129 2.17510 .153 2.068 t 0.050 2.19210 .092 2.20410 .049 2.21410 .069 2 .04710 .123 2.10510 .096 2.15210 .069 2.10310 .071 2.05010 .1 13 I .95710 .177 2.042 f0.133

(0) (0) (1) (0) (0) (0) (2) (0) (0) (0) (1) (6) (2)

" Poncent competition (% Comp) was computed by comparing test wells containing 1 Rg wbrotoxin, 5 pg AchR, and monoclonal antibody with negative control wells containing 1 pg Cobrotoxin, 5 kg AchR, and mouse preimmune serum (1 : 100). Negative control values were OD,~ = 2.08910 .035 . Positive Control wells contained 1 Kg Cobrotoxin, 5 Ng AchR, plus mouse hyperimmune serum (1 : (00) and resulted in an OD,~ value of 0.52810.026. All antibodies were tested in triplicate and the average and S.D. calculated.

Cobrotoxin Monoclonal Antibodies TABI~

3.

Antibody 2A7 1B93 6A10 15E1 13H7 13Hß 9B5 . 15F2 1892 1OD12 ID8 ßH5 IE8

hvlliBmoN

1199

oe Toxnv/AchR coun't..exxtNC (tNCUBAnoN oF ToxIN wlrll AN ANTIBODY AND AC11R è®tTUAE)

50 Ng/ml (% Comp)' 1.67210.099 1.513±0 .080 1.33510 .077 1.26210.126 1.21410 .080 L20310 .042 1.16710.253 1.09I 10.024 0.90310 .093 0.86410 .144 0.71210 .202 0.54910 .018 0.46410 .021

l0 pg/ml (% Comp) 2 ieg/ml (% Comp)

(23) (30) (39) (42) (44) (45) (46) (50) (58) (60) (67) (75) (79)

1.95410.118 1.92410.031 1.93910.142 2.07210.008 1.77010.126 1.91510.043 1 .71410.044 2.00710.013 1.51110.098 1.45210.125 1.40210.045 1.34110.101 1.13010.029

(10) (11) (11) (5) (19) (l2) (21) (8) (30) (33) (35) (38) (48)

2.21310 .025 (0) 2 .10810 .063 (3) 2.21110 .061 (0) 2.41910 .015 (0) 2.15210 .095 ( l) 2.22010 .082 (0) 2.11310 .054 (3) 2.47310 .091 (0) 2.08810 .073 (4) 2.07410 .150 (5) 1 .94810 .033 (10) 2.04710 .052 (6) 1 .88610 .015 (l3)

' Percent competition (% Comp) was computed by comparing the OD,~ readings of a 1 : 100 dilution of preimmune mouse sera (2 .17310.045) with test wells. All wells contained I ug cobrotoxin and 5 pg/AchR. Positive control wells contained 1 Pg cobrotoxin and hyperimmune mouse serum (1 : 100) which gave readings of 0.32410.019. All antibodies were tested in triplicate and the average and S.D . calculated . Epilope mapping

Peptides of eight amino acid length, moving two amino acids at a time through the entire N. n. atra cobrotoxin sequence, were made in duplicate by a pin procedure (GetrseN et ai., 1984) that is commercially available (Cambridge Research Biochemicals, Valley Stream, NY, U.S .A .) . Control peptide sequences and the synthesis schedule were generated by software from Commonwealth Serum Laboratories Commission (Victoria, Australia. Peptide-containing pins were precoated in Immulon II microtiter plates with PBS containing I ovalbumin, 1 % bovine serum albumin, and 0.1 % Tween 20 (PBSOBT) for I hr at room temperature . Monoclonal antibody was diluted to 10 ug/ml in PBSOBT . 100 pl were added per well, and incubated overnight at 4°C with peptide~ontaining pins. Pins were then washed in PBST, sheep anti-mouse antibody alkaline phosphatase conjugate (Sigma) diluted in PBSOBT was added, and incubated for I hr at room temperature . After washing with PBST, the pins were placed in a new microtiter plate containing para-nitrophenyl phosphate substrate, and the absorbance read 30 min later at 405 nm . Antibody recognition of reduced cobrotoxin was done by ELISA with cobrotoxin (10 Pg/ml in carbonate buffer, pH 9.6) that was treated with dithiothreitol (30 mM final concentration) for 6 hr at room temperature in a closed polypropylene tube . Reduced toxin and native toxin (+control) were then added to Immulon II microtiter plates ( I pgJwell) and assayed by ELISA with each monoclonal antibody (100 pl of a 10 pg/ml PBSTG solution) as previously described with mouse serum. An additional negative control consisted of incubating a I : 100 dilution of preimmune mouse serum with reduced and native toxin. All antibodies were tested in triplicate. RESULTS

Cross-reactivity of cobrotoxin monoclonal antibodies The 12 monoclonal antibodies against cobrotoxin were of the IgG isotype and, with the exception of clone 9B5 (IgGZ~, were sub-isotyped as IgG, . None of the monoclonal antibodies cross-reacted with eight other snake toxins tested by ELISA. Reactivities of the antibodies with native and reduced cobrotoxin are given in Table 1 . In vitro inhibition of cobrotoxin binding to AchR Tables 2 and 3 show the effects of monoclonal antibodies on the in vitro inhibition of cobrotoxin binding to AchR. Better inhibition of toxin binding to AchR occurred when antibody was preincubated with toxin before adding AchR (Table 2) than when antibody

120 0

B. G . STILES et ai. TABLE

4 . In vivo

Antibody Control I B93 2A7 lOD 12 9B5 1 E8 8H5

P1tOTEC1'ION SCUD>QS WITH COBROTOXIN 1lONOCLONAL ANTiHODY (300 /!g/mOUSt)

(6pg/mouse) Arm

Time to death (min)'

Time to death (test)/ time to death (control)

In vitro competition (%)

26 t 7 53 t 5 54 t 6 74 t 14 174 f 24 248 t 62 618 t 94

l .0 2 .1 2 .1 2 .8 6 .7 9 .5 23 .8

0 20 22 47 48 80 78

" The mean of the time to death was calculated from five mice used per antibody. Before injecting mice, the toxin and monoclonal antibody mixtures were preincubated together at room temperature for 30 min . Control monoclonal antibody against pseudexin did not cross-react with cobrotoxin . Mice (n = ~ injected with toxin alone had time to death readings of 33 min f 4 min . In vitro percent competition values are from Table 2 .

and AchR were incubated together with toxin (Table 3). Both tables show a dose response with decreasing antibody concentrations . The addition of 10 pg/ml monoclonal antibody (6.7 pmole/well) before 50 ~g/ml AchR (20 pmole/well) resulted in a toxin/AchR binding inhibition range of 20-80% .Incubating toxin with an AchR-antibody mixture was not as effective at similar concentrations, with toxin/AchR binding inhibition ranging from 5 to 48% . Co-incubation of five-fold more antibody (33.5 pmole antibody/well) and 20 pmole AchR/well inhibited the binding of cobrotoxin to AchR by 23-79% . This in vitro inhibition range is very similar to that seen when antibody (6.7 pmole/well) was preincubated with toxin before adding AchR (20 pmole/well). None of the monoclonal antibodies (33.5 pmole/well) added 1 hr after AchR (20 pmole/well) dissociated the cobrotoxin-AchR complex. In vivo neutralization of cobrotoxin The monoclonal antibodies delayed the time to death in mice, each given 6 hg cobrotoxin (Table 4). With a 2 : 1 mole ratio of cobrotoxin : monoclonal antibody, there was a delay in time to death of 2.3-2.7-fold compared with that of control mice given cobrotoxin plus a pseudexin monoclonal antibody that did not cross-react with cobrotoxin. A 2.5 : 1 mole ratio of monoclonal antibody : cobrotoxin yielded a 3.7-23.8-fold delay in time to death compared with the result for conrol mice (Table 4). Neutralization of toxin in mice correlated well with the in vitro binding competition values . Epitope recognition The recognition of linear epitopes by monoclonal antibodies was initially tested with cobrotoxin 8-mer peptides spanning the whole molecule but none of the antibodies bound to the toxin peptides . We obtained further evidence that the monoclonal antibodies bind to conformational epitopes by using an ELISA with reduced and native cobrotoxin (Table 1). Unlike native cobrotoxin, reduced toxin was not recognized by any of the monoclonal antibodies . Polyclonal antibodies did bind to reduced cobrotoxin (OD,°s =

Cobrotoxin Monoclonal Antibodies

12(11

0.306 f 0 .040), but to a much lesser extent than native toxin (OD, °, > 2.800). These data suggest that epitopes for the cobrotoxin monoclonal antibodies are conformational, and not linear.

DISCUSSION

This investigation characterizes 12 highly specific monoclonal antibodies against N. n . atra cobrotoxin that prevented toxin binding to AchR in vitro, and also delayed the time to death in mice lethally challenged with cobrotoxin . There have been several monoclonal antibody studies with snake venom postsynaptic neurotoùns, such as a-bungarotoxin and toxin a (BOULAIN et al ., 1982 ; DANSE et al., 1986; Z~tEINEAU et al., 1986 ; C~-tuzNC et al., 1989 ; KASE et al., 1989; PACtuvElt and RICAI,~rox, 1989). There has been only one other report of a monoclonal antibody against cobrotoxin (YANG, 1985), but little information was given and no subsequent work has been published. We found a 20-80% inhibition of cobrotoùn binding to AchR in vitro when the monoclonal antibodies were individually preincubated with toxin before the addition of AchR . Although a cobrotoxin (6 fig) : antibody mole ratio of 2 : 1 delayed the time to death in mice (2 .3-2 .7 increase in time to death compared with that of control mice), there was little variation between antibodies which gave distinctive results in in vitro competition experiments. Mice injected with cobrotoùn (6 pg = 8.6 x 10 - '° mole/mouse) and antibody (2 .0 x 10 -9 mole/mouse), a 1 : 2.5 mole ratio of toxin : antibody, showed a more marked effect with a 3 .7-23.8-fold delay in the time to death compared with that of control mice . The in vitro and in vivo (1 : 2 .5 mole ratio of toxin : antibody) protection studies correlated well . The antibodies that best prevented the binding of toxin to AchR in vitro also afforded the longest time to death in mice. Antibodies with little effect on in vitro binding of toxin to AchR concomitantly yielded shorter times to death in mice. There was some variation noted between the in vitro and in vivo assays. Antibodies 8H5 and 1E8, which had nearly identical in vitro inhibition values (78 and 80%, respectively) for toxin binding to AchR, differed by 2.5-fold in in vivo neutralizing capabilities. Another example of monoclonal antibodies with similar in vitro competition values is 1OD12 and 9B5 (47 and 48%, respectively) . The in vivo results for 1OD12 and 9B5 differed by 2.4-fold. Antibodies with low in vitro inhibition values, such as 2A7 and 1 B93 (22 and 20%, respectively), provided identical in vivo results. Variations between assays with antibodies that share similar in vitro competition values may be attributed to the affinity of an antibody for toxin in solution vs toxin adsorbed to an ELISA plate. There was little correlation between antibody reactivity with cobrotoxin and inhibition of toùn binding to AchR in vitro (Tables 1 and 2) . The most remarkable differences were the relatively low reactivity of antibody lE8 with cobrotoxin (OD,°S = 0.905±0 .013) but high toùn/AchR binding inhibition in vitro (80%). On the other extreme was antibody 15E1 and its higher reactivity with cobrotoxin (OD,°S = 1 .707±0 .032) yet lower toxin/ AchR binding inhibition in vitro (21 %) . A previous in vivo neutralization study with N. nigricollis toxin a (BOULAIN et al., 1982) and a monoclonal antibody showed that a 15 .5 : 1 mole ratio of antibody :toxin (3 hg toùn/mouse) protected all mice (n = 10) for 24 hr. Since the study was not done for a longer time period, it remains uncertain if the mice were protected over an extended time period. The monoclonal antibody recognized a site reportedly not involved in AchR binding but protection was probably due to steric hindrance of toùn binding to AchR

120 2

B . G . STILES et a/.

and/or the toùn was conformationally changed such that the affinity of the toxin for AchR was decreased. Dissociation of the toxin a : AchR complex was achieved with a 900 molar excess of monoclonal antibody vs toxin in a radiolabeled toxin assay (BOULAIN et al., 1985). Our in vitro studies with a 13-fold molar excess of any of the cobrotoxin monoclonal antibodies, compared with plate-adsorbed cobrotoxin able to bind AchR (STtLFS, 1991), did not dissociate the toxin/AchR complex. Another monoclonal antibody/toxin a report by Ttt>a~nu et al. (1986) describes a different monoclonal antibody which recognized and neutralized various short~hain neurotoxins in vitro, but only delayed mouse lethality, even at a 50 : 1 mole ratio of antibody : toxin a (MEtvEZ, 1985). The neutralizing epitope was evidently conformational and incorporated at least the invariant amino acids Lys-27, Trp-29, and Lys-47 ('IytE~nu et al., 1986). We did not try in vivo protection experiments with such a high mole ratio of antibody :toxin . In vivo neutralization studies with monoclonal antibody and the long-chain neurotoxin, a-bungarotoxin, also show only a delay in the time to death in mice (DnxsE et al., 1986; KnsE et al., 1989). The inability of monoclonal antibody to prevent lethality may be due to the higher affinity of toxin for AchR than antibody and/or the processing and excretion of antibody may be faster than toxin (DnNSE et al., 1986). DnxsE et al. (1986) found that mole ratios of 10 : 1 (antibody : a-bungarotoxin), using 10 pg toxin/mouse with each of the 26 monoclonal antibodies, increased the time to death 6-100-fold compared with the results for control mice given toxin alone. Another study with mole ratios of 1 : 1 and 5 : 1 of monoclonal antibody : a-bungarotoxin (10 hg toxin/mouse) increased the time to death 4 and 13 .5-fold compared with that of control mice given toxin alone, respectively (KnsE et al., 1989). KnsE et al. (1989) found that the same monoclonal antibody against a-bungarotoxin also recognized other long-chain neurotoxins. Unlike studies that involved the short-chain neurotoxin, toxin a, linear neutralizing epitopes on a-bungarotoxin (amino acids 34-41 and 34-38) were recognized by two different monoclonal antibodies (CxunxG et al., 1989 ; Kris et al., 1989). Our investigation suggests that the monoclonal antibodies against cobrotoxin recognized conformational epitopes as none of the monoclonal antibodies reacted with reduced cobrotoxin or synthesized peptides spanning the whole toxin molecule. Since polyclonal antibody bound to reduced cobrotoxin, albeit much less effectively than native toùn, there was some antibody recognition of linear sequences or all the toxin molecules were not thoroughly reduced. Besides the recognition of conformational epitopes by monoclonal antibodies against N. nigricollis toxin a, polyclonal antibodies against a cardiotoxin from N. nivea venom also bind exclusively to conformational epitopes (05THOFF, 1989). This is of interest because some cardiotoxins and short-chain postsynaptic neurotoxins are strikingly similar in isoe~ectric point, amino acid length, conformation, and number of disulfide bridges (VissEx and Louw, 1978). As there was little reactivity of cobrotoxin monoclonal antibodies with other snake toxins, including N. n. atra cardiotoxin, the epitopes must be ~~~.iique to cobrotoxin . Lack of reactivity by the monoclonal antibodies with the tested toxins further suggests that linear epitopes were not recognized because many snake venom postsynaptic neurotoxins are very homologous (Dur~-rox and Hrosx, 1983). Experiments are planned to determine the epitopes bound by the cobrotoxin monoclonal antibodies. All monoclonal antibody studies to date with snake venom postsynaptic neurotoxins show that a large molar excess of antibody is required for an in vivo protective effect,

Cobrotoxin Monoclonal Antibodies

1203

although transient in many cases. Our study also demonstrates a transient, yet not absolute, protective effect with monoclonal antibodies against cobrotoxin . The nonradioactive receptor assay provided some clues on the effectiveness of different antibodies in preventing the binding of toxin to AchR in vivo. Furthermore, development of a truly efficacious immunotherapy/vaccine against various short- and long-chain postsynaptic neurotoxins found in snake venoms will most likely involve more epitope mapping studies with monoclonal antibodies. REFERENCES AaE, T. and TAMIYA, N. (1979) Immunological studies on erabutoxin b, a sea snake toxin. Attempts to locate the amino acid residues determining antigenicity . Toxicon 17, 571-582. BoQuET, P., PoILLEUx, G., DUMAREY, C., IZARD, Y. and RoNSSPxAY, A.-M. (1973) An attempt to classify the toxic proteins of Elapidae and Hydrophidae venoms . Toxicon 11, 333-340. BOULAIN, J.-C., MENEZ, A., COUDERC, J., FAURE, G., LIACOeouLOS, P. and FROMAGEOT, P. (1982) Neutralizing monoclonal antibody specific for Naja nigricollis toxin a: Preparation, characterization, and localization of the antigenic binding site. Biochemistry 21, 2910-2915. BOULAIN, l .-C., FROMAGEOT, P. and MENEZ, A. (1985) Further evidence showing that neurotoxin-acetykholine receptor dissociation is accelerated by monoclonal neurotoxin-specific immunoglobulin . Molec. Immun. 22, 553-556. C~IANG, C. C. and YANG, C. C. (1969) Immunochemical studies on cobrotoxin. J. Immun. 102, 1437-1444. CwANG, C. C. and YANG, C. C. (1973) Immunochemical studies on the tryptophan-modified cobrotoxin . Biochim. biophys. Acta 295, 59504. CSiUANG, L. Y., LIN, S. R., CHANG, S. F. and GANG, C. C. (1989) Preparation and characterization of monoclonal antibody specific for a-bungarotoxin and localization of the epitope . Toxicon 27, 211-219. DArtsE, J. M., ToussAlrrr, J. L. and KEMPF, J. (1986) Neutralization of a-bungarotoxin by monoclonal antibodies . Toxicon 24, 141-151. DuFroN, M. J. and HIRER, R. C. (1983) Conformational properties of the neurotoxins and cytotoxins isolated from elapid snake venoms. CRC Crit . Rev. Biochem. 14, 113-171 . EARLEY, E. M. and OSfF.RLING, M. C. (1985) Fusion of mouse-mouse cells to produce hybridomas secreting monoclonal antibody . J. Tiss . Cult. Meth . 9, 141-146. GEYSEx, H. M., MELOEN, R. H. and BARTELIxG, S. J. (1984) Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid . Proc . natn. Acad. Sci. U.S .A . 81, 3998-4002. JUILLERAT, M. A., SCHWENDIMANN, B., HAYERT, J., FULPIUS, B. W. and BARGEfZI, J. P. (1982) Specific binding to isolated acetykholine receptor of a synthetic peptide duplicating the sequence of the presumed active center of a lethal toxin from snake venom. J. biol. Chem . 257, 2901-2907. KASE, R., KITAGAWA, H., HAYASHI, K., TANGUE, K. and INAGAKI, F. (1989) Neutralizing monoclonal antibody specific for a-bungarotoxin : Preparation and characterization of the antibody, and localization of the antigenic region of a-bungarotoxin . FEBS Lett . 254, 106-110. Low, B. W. and CoRF1Ein, P. W. (1986) Erabutoxin b: Structure/function relationships following initial protein refinement at 0.140 -nm resolution. Eur. J. Biochem. 161, 579-587. McDANIEL, C. S., MANSHOURI, T. and ATAS31, M. Z. (1987) A novel peptide mimicking the interaction of a neurotoxins with acetykholine receptor . J. Prot. Chem . 6, 455-461 . MENEZ, A. (1985) Molecular immunology of snake toxins . Pharmac. Thcr. 30, 91-113 . MENEZ, A., BOULAIN, J. C. and FROMAGEOT, P. (1979) Attempts to define the antigenic structure of Naja nigricollts toxin a. Toxicon 17, 123. Music-JERICEVIC, B. and ATAS31, M. Z. (1987) Profile of the a-bungarotoxin binding regions on the extracellular part of the a chain of Torpedo calijornica acetykholine receptor . Bicehem. J. 248, 847-852. Music-JEwcevlc, B., MANSxouRl, T., Yoxtx, T. and ATA39I, M. Z. (1988) The regions of a neurotoxin binding on the extracellular part of the a subunit of human acetykholine receptor. J. Prot. Chem. 7, 173-177. OSTHOFF, G. (1989) Far-u.v. CD-spectroscopy and immunological properties of synthetic sequential peptides derived from cardiotoxin V"1 of Naja nivea venom: An amphipathic a-helix formed by sequence IS-25 of a ß-protein. Int. J. Biochem. 21, 1365-1368 . PACHNER, A. R. and RICALTON, N. (1989) In vitro neutralization by monoclonal antibodies of a-bungarotoxin binding to acetykholine receptor. Toxicon 27, 1263-1268 . RENER, J. C., BROWN, B. B. and NARDONE, R. M. (1985) Cloning hybridoma cells by limiting dilution . J. Tics. Cult . Meth . 9, 175-177. RUAN, K:H., $PURLING, J., Quloclto, F. A. and ATA33I, M. Z. (1990) Acetylcholine receptor-a-bungarotoxin interactions : Determination of the region-to-rogion contacts by peptide-peptide interactions and molecular modeling of the receptor cavity. Proc. natn . Acad. Sci. U.S.A . 87, 6156-6160.

1204

B. G. STILF.S ct d.

Sxutatwx, M., Wtt.n>~, C. D. and Ko~, G. (1978) A better cell line for making hybridonas secreting spaific antibodies. NoteQe 276, 269-270. STIIL'4, B. G. (1991) A non-radioactive receptor assay for snake venom postsynaptic neurotoxins. Toxicon 29, 503-510. Tte~au, O., Houunv, J.-C., Couneac, J., FROSUC®or, P. and MENBZ, A. (1986) A monoclonal antibody which reoogetized the functional site of snake neurotoxina and which neutralizes all short-chain variants. FEBS Lctt. 208, 23lr240. Vts~x, L. and Louw, A. I. (1978) The conformation of cardiotoxins and neurotoxins from snake venoms. Biochim . biophys . Acta 533, 80-89. WALICINSHAW, M. D., Seeeveea, W. and MAELIC[CE, A. (1980) Three-dimensional structure of the "long" neurotoxin from cobra venom . Proc . natn. Acad. Sci . U.S .A . 77, 2400-2404. Wept, M. and C~texaeux, J:P. (1974) Binding of Naja nigricollù (~H]a-toxin to membrane fragments from Electrophorus and Torpedo electric organs . I. Binding of the tritiated a-neurotoxin in the absence of effector. Molec. Pharmac. 10, 1-14. YANG, C. C. (1965) Crystallvation and properties of cobrotoxin from Formosan cobra venom. J. bio(. Chore . 210, 1616-1618. Yt tvc, C. C. (1974) Chemistry and evolution of toxins in snake venoms . Toxicon 12, 1~3. YANG, C. C. (1985) Immunochemical studies on cobrotoxin . Toxicon 23, 629. Yf,NG, C. C., Yexc, H. J. and Hu~rtc, J. S. (1969) The amino acid sequence of cobrotoxin . Biochim . biophys . Acts 188, 65-77. YANG, C. C., Y~rtc, H. J. and G~tu, R. H. (1970) The position of disulûde bonds in cobrotoxin . Biochim . biophys. Acts 214, 355-363. Yexc, C. C., Lrx, M. F. and C~rc, C. C. (1977) Purification of anti-cobrotoxin antibody by amity chromatography. Toxicon 15, 51-62. Y~rrc, C. C., CFUtve, C. C., Lnv, M. F., LEA, H. Y. and Cttunxc, L. Y. (1978) Purification and properties of non-precipitating antibodies to cobrotoxin . Int . J. Peptide Protein Res. 12, 303-310.