Food Research Institute, 1925 Willow Dr., University o f Wisconsin, Madison, WI53706, U.S.A.
(Received 29-10-1987, accepted after revision 27-1#988)
Summary m The type B neuro toxin (NT) isolated ti 'om Clostridium botulinum (strain 657) behaved as a mixture of single (unnicked) and dichain (nicked) prote ins , both of M. ~1 5 0 kDa. When the dichain NT was r educed by mercap toe thano l , the two chains migra ted in sodium dodecyl su l fa te -po lyacry l - amide gel e lect rophores is ( S D S - P A G E ) as separate polypept ides of Mr ~ 100 and 50 kDa that appear - ed similar to the heavy and light chains of o the r sero types o f botul inum NT. T h e N-terminal amino acid sequences of the two chains were de te rmined . Th ey were as follows: light chain: P r o - V a l - T h r - l l e - A s n - A s n - Phe - A s n - Tyr - Ash - Asp - Pro - Ile - Asp - A s n - A s n - Asn - Ile -- Ile - Met - Met - Glu - Pro - P r o - Phe - Ala - Arg - Gly - Met - Gly - Arg - Tyr - Tyr - Lys - Ala - Phe - Lys - Ile - Th r - Asp - Arg - Ile - T rp - Ile - ; and heavy chain: Ala - Pro - Gly - Ile - X - Ile - A s p - Val - Asp - A s n - Glu - Asp - Leu - Phe - Phe - lie - Ala - Asp - Lys - A s n - Ser - Phe - Arg - Asp - Asp - Leu - . These two sequences matched exactly with those of the light and heavy chains of type B N T (strain Okra ) of which only 16 and 18 residues were known (J. Biol. Chem. (1985) 260, 10461). The above sequences were dif ferent f rom those o f type A NT. lmmunoprec ip i ta t ion react ions of type B N T isolated f rom strains 657 and Okra were indistinguish- able against polyclonal anti-type B N T serum. These two prepara t ions did not p roduce precipit in reac- tions with polyclonal anti-type A N T serum. The te tanus N T (f rom C. tetani) resembles botulint~m NT in relative molecular mass, dichain s tructure and probab ly in the mode of act ion; but the two NTs differ in their sites of action and their serological specificity. Compar i son of amino acid sequences of botulin- um NT, r epor t ed here , and te tanus N T (EMBO J. (1986) 5, 2495) shows that thei r two heavy chains are less homologous (13 out of 26 residues matched in posi t ion) than the two light chains (31 out o f 44 residues matched) . Three stretches o f residues: 3 - 7 , 1 8 - 2 4 and 3 2 - 4 4 , on the light chains of the two NTs are identical.
botulinum type B neurotoxin / tetanus neurotoxin / sequence / homology
Introduction
The anaerobic bacter ium, Clostridium botuli- num, produces the protein botul inum neuro- toxin (NT) , Mr ~150 000, in antigenicaily dis- tinct forms ( types A, B, Ct, D, E, F and G). One strain general ly produces one sero type , yet there are except ions [1]. The strain 657, isolated f rom a case of human infant botul inum, produced type B toxin-that appeared unusual in its serolo- gical proper t ies [2]: " O n e internat ional unit of type B anti toxin neutral ized only abou t 10 lethal doses of 657 toxin as compared with approximate-
ly 10 000 lethal doses of convent ional type B toxin f rom the Bean s t ra in" . Anti- toxin prepar- ed against 657 toxin was 10 times more effective against the convent ional toxin than against the homologous toxin. H a t h e w a y et al. [2] enter ta in- ed two possible explanat ions: 1) The bacter ia (strain 657) p roduce a toxic and a non-toxic form of a prote in possessing similar antigenic / immu- nologic epi topes and the re fo re both forms com- bine with anti- type B serum. 2) The organism produces " m o r e than one serological s p e o e s o f toxic prote in and that only one cor responds to the type B specificity".
8i2 B. R. DasGupta and A. Datta
The latter idea was extended by Gimrnez [3] who reported that the 657 "toxin is a mixture of B (approximately 95% of the complex) and A anti- genic fractions" and the toxin from strain 657 represents the prototype of a new serotype named subtype Ba. This conclusion was based on cross-neutralization tests done with anti- serum prepared against NT serotypes A, B and B 657. Production of type A and B NTs, as a mixture, by a strain of C. botulinum (I.P. 7212) has been documented [4].
Because the partial amino acid sequences of type B (strain Okra, i.e., Bean) and type A (strain Hall) were already known [5, 6] and anti- sera against pure type A N T and pure type B NT were available, we designed experiments to compare the 657 toxin with type A and type B (strain Okra) NTs on the basis of amino acid sequence and serological reactivity.
The p,-ocedures used for the purification of type A [7, 8] and type B [9-11] NT are different; interchanging the procedures does not produce a pure NT type. We purified the 657 toxin fol- lowing a method of purification of type A N T , rather than that of type B (Okra) NT, with the hope that purified 657 toxin could be enriched in type A N T . Following its purification, the 657 toxin was identified serologicaUy by testing against anti-type A N T and anti-type B NT serum. Also, the two chains of the 657 toxin were separated and analyzed for amino acid sequence.
The tetanus ""~ duced by . . . . . . L l O S I r l t l l U m i ' l l , p¢o tetani, is serologically distinct from botulinum NT. They are similar in relative molecular mass, dichain structure and in their mode of action but not in their specific sites of action [12, 13]. We compared the sequences of the two chains of the type B 657 toxin, reported here, with those of the published sequences of the two chains of tetanus NT [14].
Materials and methods
Culture The stock culture of C. botulinum (strain B657) ini- tially received from Dr. C. Hatheway (Centers for Disease Control, Atlanta, GA) was further examin- ed by Dr. L. Siegal (Fort Detrick, MD) based on streaking plates and picking a single colony. This stock culture was incubated in a cooked meat medium (Difco Laboratories, Detroit, MI) at 37oC for 24 h. 1 ml of the growing culture (fluid from the cooked meat media) was transferred into a 250 ml toxin production medium (see below) for incubation
at 37°C for 24 h. This actively growing culture was poured into 15 I of toxin production medium. After 6 days of incubation at 37oC, this culture (pH had dropped to 5.8 from the initial pH 7.0) was the source of toxin. Toxin production medium [10]: a solution of 2% N-Z Amine type B (Sheffield Chem.) and 1.5% yeast extract (Difco) adjusted to pH 7.0 with 5 N NaOH was autoclaved; glucose, autoclaved sepa- rately, was added to the medium to a 1% final concentration.
Purification The 6 day old culture acidified with 3 N H2SO4 to pH 4.0 was stored in a cold room (8oC) for at least over- night. The precipitate was centrifuged (19 000 × g, 20 min, 4oC), washed twice with distilled water and then extracted for 1 h at 25oC with --100 ml of 0.2 M NaH2PO4-NazHPO4 buffer, pH 6.0 (containing 1.7/zg aprotenin/ml, Sigma). The extract was packed (23 000 x g, 20 min, 4oC) and the pellet re- extracted with pH 6.0 phosphate buffer twice as above. The supernatants were pooled (total -320 ml), partially saturated with solid ammonium sulfate (31.3 g/100 ml) and held at 8oC for at least 48 h. The precipitate, recovered by centrifugation (12 100 x g, 30 min, 4oC) was dissolved in 0.2 M sodium phos- phate buffer, pH 6.0, (total vol. --110 ml), and then digested with ribonuclease (EC 2.7.7.16, ribonu- clease A, CalBiochem lot #556746, 50/zg/ml) at 30°C for 5 h. The digest was partially saturated with ammonium sulfate (31.3 g/ml) and stored at 8oC. The precipitate, recovered by centrifugation (30 900 × g, 30 rain, 4oC), was dissolved in --20 ml 0.05 M sodium citrate buffer, pH 5.5, and then dialyzed against 500_ml of pH 5.5 buffer for 3 h at 25°C, chang- mg the butter every h. The dialyzed material was chromatographed on DEAE-Sephadex A-50 columns as described earlier [7]. Fractions across the first peak with an A26o/A278 ratio up to 0.6 eluted from the two columns were combined (total ~96 ml; 1.91 A260, 3.12 A278). Further steps for chromatogra- phy on Sephadex G-100 columns were according to [7]. Only one peak emerged even after elution with 200 ml of buffer. Fractions of this peak from the two parallel columns were combined (total - 100 ml; 1.32 A260, 2.29 A278). Concentration of the protein and steps taken for chromatography on DEAE-Sepha- dex A-50 columns at pH 7.9 were described [7]. The fractions of the first major peak eluted under gra- dient (Fig. 1) were pooled (from the two parallel columns), dialyzed in a cold room (48 h) against 0.02 M sodium phosphate buffer, pH 5.8, and then for 2 h at room temperature and finally chromato- graphed on an SP-Sephadex C-50 column (Fig. 2). Several fractions across the main peak were analyzed by SDS-PAGE [15]. Details of the electrophoresis were reported [16]. One pan of the peak was used for amino acid sequence determination and another part was precipitated with ammonium sulfate (31. ! g/100 ml) for use in se/-ological assays.
Botulinum neurotoxin type B (strain 657) 813
a~
L92 -
l .O0 -
0.50 -
' W :f ' - , /
I0
G 1
I I I , I I
20 30 4 0 50 60 70 80 90 lO0
Tube Number
Fig. 1, Chromatography of the B657 toxin on a DEAE-Sephadex A-50 column (1.5 × 20 cm) equilibrated with 0.02 M NaH2POa-Na2HPO4 buffer, pH 7.9, at 25oC. Sample load: -52 ml of 2.69 A27 s and 1.64 A2~. Fraction size, 4 ml/ tube. Arrows W and G, mark the beginnings of column wash (tube 13) and linear gradient elution at tube 24 (150 ml equilibrating buffer plus 150 ml equilibrating buffer containing 0.3 M NaCI). The toxin was recovered in tubes 58-68; pooled volume 44 ml, 0.63 A+Ta; A278/Aat~t = 0.52.
1.00
CO
O d ,~ 0.50
I I . -
I 0 2 0 $0 4 0 5 0 6 0 7 0 8 0 9 0 I O0
Tube Number
Fig. 2. Chromatography of the B657 toxin on an SP-Sephadex C+50 column (1.5 × 20 cm) equilibrated with 0.02 M NaH2PO4-Na2HPO4 buffer, pH 5.8, at 25°C. A pool of the fractions from DEAE-Sephadex columns (tvbes 58-68 in Fig. 1, and similar fractions from a parallel column), following dialysis against pH 5.8 buffer was loaded; volume 120 ml, 0.287 A278 and 0.172 A260. Size of fractions up to tube 50:4 ml / tube; then 3.5-ml / tube. Arrows W and G, mark the beginnings of column wash (tube 31) and linear gradient elution at tube 5l (100 ml equilibrating buffer plus 100 ml equilibrating buffer containing 0.3 M NaCl). Fractions 77, 78, 81,86, 87 and 96 were analyzed by SDS- PAGE (7.5% gels, 6 h, 8 mA / gel). See inset. Gel 1 (sample from fraction 87, not reduced): NT at 9-11 mm; gel 2 (sample from fraction 87, reduced) heavy chain at 14-16 ram, light chain at 30-32 ram), other bands in between are impurities. Electrophoretic migration from left to right.
814 B. R. DasC,,tpta and A. Datta
Trypsinization The purified toxin preparation was incubated with trypsin (EC 3.4.4.4, tosylamido-2-phenylethyl- chloromethyl ketone-treated, Cooper) at 30:1 (w/w) ratio in 0.1 M sodium acetate buffer, pH 6.0, for 30 min at 37"C.
Amino acid sequence The heavy and light chains of the NT were separated by SDS-PAGE (see Fig. 2 inset, gel 2). Coomassie blue-stained gels were sliced with a razor blade and the stained protein was recovered from the gel by electroelution [ 17]. The amino acid sequences of the two chains were determined using a gas-phase sequencer (Applied Biosystems Moel 470A, see [18]). The phenylthiohydantoin (PTH) derivatives were identified using reverse-phase high pressure liquid chromatography (HPLC) (IBM instruments LC/9533 and CS 9000 [ 19]) by comparing the chro- matograms of each cycle to that of an internal stan- dard PTH-norvaline. The criteria for designating the nth residue in the polypeptide required the most abundant PTH-amino acid to show an increase over the previous cycle (n-1 residue) and decreases in subsequent (i.e., n + 1, n + 2) cycles.
lmmunochemical techniques Serological reactivity (Ouchterlony gel double diffu- sion) was assayed in plates made with 1.5% Noble- Agar (Difco Laboratories) in 0.05 M sodium phos- phate buffer, pH 7.5, containing 0.85% NaCI. The rabbit anti-type A N T serum (no. 133), which has been used before [20, 21], and anti-type B serum (no. 200) were prepared against the respective pure NTs that were detoxified with formalin. For immuno- chemical comparison, oure tvoe A N T , oroduced from a cultureof the Hall strain; and type B'NT, pro- duced from a culture of the Okra strain, were purifi- ed in our laboratory [7, 11].
Results and Discussion
Purity and characterization
The toxin produced by C. botulinum type B (strain 657) culture was significantly purified after S P - Sephadex chromatography (total yield ~6.8 mg of protein from 15 1 culture, see [22] for estimation of protein content) but it contained trace impurities (see inset in Fig. 2 for SDS- PAGE gels). Note: This purification procedure, which was developed for type A N T [7], yields pure type A [7, 22], but not pure type B NT (see Introduction). The protein (not reduced with mercaptoethanol) migrated electrophoretically as a major band (Mr 150 000); there were faint bands ahead of the main band (Fig. 2 inset). In
the mercaptoethanol-reduced sample, the major band (Mr ~150 000) was replaced by two strong bands of Mr ~ 100 000 and ~50 000, but a trace of the Mr 150 000 band remained (unnicked sin- gle chain protein). The reduced sample also show- ed impurities. These results demonstrated that the toxin, significantly purified, was: 1) a mix- ture of a dichain protein and a single chain pro- tein, both of similar Mr --150 000; and 2) the dichain protein was composed of two poly- peptide chains joined by disulfide bond(s). Mild trypsinization of the sample did not alter the electrophoretic behavior of the Mr 150 000 band. When the trypsinized sample was reduced with mercaptoethanol, the small amount of the Mr 150 000 band (unnicked single chain protein) was eliminated (not shown here in gels). Clearly, a small amount of the single chain (unnicked) toxin present in the mixture of single and dichain protein was converted (nicked) to the dichain protein by trypsin.
The above observations are consistent with one known feature of botulinum NT [12, 16]. The NT isolated from bacterial culture is a single chain, a dichain or a mixture of single and dichain proteins. Age of the cukure and physiol- ogy of the bacteria determine the relative propor- tion of the single and dichain NT. In the bacterial culture, the single chain NT (Mr 150 000) is nicked to the dichain molecule by an endo- genous protease(s). The two chains (heavy and light, Mr ~100 000 and ~50 000, respectively) remain held . . . . . ~'^- - S S k__.~o ,,~:,.~ trypsin treatment also nicks the single chain NT into the dichain form.
Imrnunochemis try
The toxin produced by strain 657 did not react with anti-type A NT serum (Fig. 3, left plate). It reacted with anti-type B NT serum (Fig. 3, right plate): the precipitin lines that formed between anti-type B serum and type B NT (strain Okra),well 1, and the toxin from strain 657, well 2, were indistinguishable. The anti-type B serum did not react with type A N T . The anti-type A NT serum produced strong precipitin lines against type A NT, whether native or treated with a tryptophan-specific reagent (2-hydroxy-5- nitrobenzyl bromide). Further evidence of the specifici-:y of this anti-type A NT,~rum (no. 133) was that it did not reac~ with the type A hemag- glutinin (well 6, left plate). This hemagglutinin remains tightly complexed with the NT in the bacterial culture fluid and ihrough several steps
Botulinum neurotoxin type B (strain 657) 815
Fig. 3. Serological reaction (Ouchteriony gel diffusion) of type B657 toxin. Left plate: central well: anti-type A NT serum; wells I and 4: type A N T (treated with 2-hydroxy-5-nitrobenzyl bromide); well 2: type A NT (native); wells 3 and 5: type B657 toxin; well 6: hemagglutinin protein from type A. Right plate: central well: anti-type B NT serum; well 1: type B NT (from strain Okra); wells 2 and 4: type B657 toxin; wells 3 and 6: type A N T (native and 2-hydroxy-5-nitrobenzyl bromide-treated, respective- ly); well 6: hemagglutinin from type A. Anti-type A and B sera were prepared with immunogen detoxified with formalin.
of purification [7, 12}. Also this hemagglut inin cross-reacts immunological ly with a hemaggluti- nin in type B cultures [23].
Sequence
The sequence of the first 44 amino acid residues of the light chain f rom B657 toxin was found to be as follows: Pro - Val - T h r - Ile - Ash - Asn - Phe - A s n - Tyr - A s n - Asp - Pro - Ile - Asp - A s n - Ash - Asn - Ile - l ie - Met - Met - G |u - Pro - Pro - Phe - Ala - Arg - Gly - Met - Gly - Arg - Tyr - Tyr - Lys - Ala - Phe - Lys - l ie - T h r - Asp - Arg - Ile - T rp - Ile - . The sequence of the first 26 amino acid residues of the heavy chain was found to be as follows: Ala - Pro - Gly - l le - X - Ile - Asp - Val - Asp - A s n - Glu - Asp - Leu - Phe - Phe - Ile - Ala - Asp - Lys - Asn - Set - Phe - Arg - Asp - Asp - Leu - . The fifth residue of the heavy chain (marked X) could not be identified by H P L C because the sample was not ca rboxymethy la ted . Two independen t prepara t ions of the B657 toxin were analyzed by S D S - P A G E . The heavy and light chains f rom these two prepara t ions were sliced out of gels and were analyzed for se- c~uences. T h e sequences of similar chains f rom
fllatt;llCU. I t ic two independen t prepara t ions . . . . . . . sequences of only 16 and 18 residues of the light and heavy chains of type B N T (strain Okra) were known [6] and they matched completely with the present results (Fig. 4). The identi ty of the X, the fifth residue of the heavy chain, we
inferred to be half cystine because that is the fifth residue of the heavy chain in type B N T [6].
T h e toxin isolated f rom strain 657 appeared to be identical to the type B N T (strain Okra) by these two criteria: 1) Serological reactivity o f this toxin prepara t ion was indistinguishable f rom type B N T (strain Okra) and was not reac- tive to the ant i - type A N T serum. 2) T h e partial amino acid sequences of the light and heavy chains of 657 toxin were identical to the partial amigo acid sequences of the light and heavy, chains of type B N T (strain Okra) . Compar i son of the por t ions of the light and heavy chains sequenced so far with the corresponding se- quenced por t ions of type A N T [5, 24] (Fig. 4) fu r ther co r ro b o ra t ed the conclusion that the toxin f rom 657 is type B and not type A. A cer- tain amount o f sequence homology among botu- l inum N T sero types A and B is expec ted because they are pharmacological ly similar; this was demons t ra ted am o n g types A, B and E [22]. In spite of this background , the a rgument that s tands out f rom Fig. 4 is that the amino acid sequences of the neuro toxic prote in isolated f rom stain 657 and Okra are comple te ly identical and both differ identically f rom the sequence of type A NT.
Results of sequence work, in addit ion to the serological react ion, demons t ra ted that the 657 toxin we p r ep a red did not have a detectable am o u n t of type A N T . Residues 1, 7 - 9 , 11 and 12 of the light chains of types A and B N T are identical, while residues 2 - 6 and 1 3 - 1 6 are not. If the 657 toxin prepara t ion conta ined a detect-
816 B. R. DasGupta and A. Datta
Light chain
Type A Type B (Okra) Type B (657) Tetanus
l IO 20 30 40 P.F.V.N.K.Q.F.N.Y.K.D.P.V.N.G.V.D.- P.V.T.I.N.N.F.N.Y.N.D.P.I.D.N.N.- P•V•T•I•N•N•F•N•Y•N•D•P•I•D•N•N•N•I•I•M•M•E•P•P•F•A•R•G•M•G•R•Y•Y•K•A•F•K•I•T•D•R•I•W•I•- P•I•T•I•N•N•F•R•Y•S•D•••v•N•N•D•T•I•I•M•M•E•••P•Y•C•K•G•L•D•I•Y•Y•K•A•F•K•I•T•D•R•I•W•I•-
Heavj/ chain
Type A Type B (Okra) Type B (657) Tetanus
l IO 20 A.L.N.D.L.C.I.K.V.N.N.W.D.L.F.F.S.P.S.E.D.N.F.T.N.D.L.°
Fig. 4. Partial amino acid sequences of the light and heavy chains of the type B toxin (from strain 657), compared with botulinum NT types A and B (from strain Okra) and tetanus NT. Amino acid residues underlined are homologous between the tetanus and botulinum type B NTs. Sequences of type A (5 and 24) and type B (6) (from strain Okra) botulinum NT were reported before. Sequences of tetanus NT were derived from nucleotide sequence by Eisel et al. [14]. The fifth residue of the heavy chain (marked X) could not be identified by HPLC; the sample was not carboxymethylated.
able amount (as much as 5%) of type A N T , the sequence runs of the light chain of 657 toxin would have revealed, at cycles 2 - 6 and 13-16, the residues of type A N T . Two separate prepar- ations of the light chain sequenced at 55 and 200 pmol did not show any of the diagnostic residues of type A NT.
The repo~ted serological studies of the 657 toxin (neutralization of mouse lethality) used impure material; bacterial culture supernatant 131 and precipitate obtained from the whole culture with acid or ammonium sulfate [2] were used. Hence, these extremely complex mixtures could have contained "more than one serological species of toxic protein' [2] or 'antigenic frac- tions' [3]. Goal of the present study was to char- acterize, chemically and immunochemically, the NT named subtype Ba. Thus the reported sero- logical behavior of the very complex mixture of materials produced by C. botulinum strain 657, cannot be explained now, other than: 1) the NT is identical to type B NT, and 2) if a small amount of type A N T was also produced by the organism, it was ~emoved during purification of the type B NT.
Comparison with tetanus N T
Botulinum and tetanus NTs have 3-step mecha- nisms of action that are similar [13, 25]: 1) the NTs bind to their respective receptors on a spe- cific presynaptic membrane, 2) the NTs, or some part of the NTs, translocate, and 3) then inside the cell, the translocated NTs (or part of the
NTs) produce a metabolic lesion that blocks neurotransmitter release. Although their gener- al modes of action are similar~ their sites of action differ [13]; tetanus NT at very high concentration recognizes the receptors of botuli- num NT [26]. One of us (BRDG) had pointed out in 1977 [12], for the first time, the structural similarities between botulinum NT and tetanus NT, such as their relative molecular mass and diehain structure. Since then, it has become known that tetanus and botulinum NTs have cer- tain similarities in amino acid_ compositions [22]° Very limited amino acid sequence data of botuli- num NT also revealed some homology with teta- nus NT [6, 22]. The 70 residues (44 residues from the light chain plus 26 residues from the heavy chain) of the botulinum NT strain 657, which appears to be type B NT, sequenced and report- ed here are the most we know of amino acid sequences of any botulinum NT serotype. These sequences were compared with those of tetanus NT (Fig. 4). The sequences of the light chains of the type B and tetanus NT [14] have 70% homol- ogy (31 out of 44 residues matched); the two heavy chains are less homologous (13 out of 26 residues matched in position). Among the light chains of the two NTs, three stretches of residues 3 - 7 , 18-24, and 32-44, are identical. The two heavy chains share three homologous stretches, none of which is longer than four residues.
The mode of action and s tructure-funct ion relationship of the two dichain NTs [13, 16, 27] suggest that, for each NT, the role of one chain (heavy) is receptor selection and binding, where-
Botulinum neurotoxin type B (strain 657) 817
as intracellular metabolic lesion is caused by the other chain (light?). This allows one to conjec- ture that the heavy chains of the botulilnum and tetanus NTs are likely to differ structurally (they act on different nerve cells) more than the light chains (which cause similar intracellular meta- bolic lesions?). The available sequence data, 50% homology between the heavy chains and 70% homology between the light chains, are consistent with the conjecture.
Acknowledgments
Our thanks to M. A. Woody, J. Foley Jr. and R. Niece for their assistance in serological reactions, gel slicing and amino acid sequence work. This study was supported in part by the United States Army Medical Research and Development Command (DAMD 17- 84-C-4245), National Institutes" of Health grant NS 17742, the Food Research Institute and College of Agricultural and Life Sciences.
References
1 Sakaguchi G. (1983) Pharmacol. Ther. 19, 165-194
2 Hatheway C. L., McCroskey L. M., Lombard G. L. & Dowell V. R. Jr. (1981) J. Clin. Micro- biol. 14, 607-611
3 Gim6nez D. F. (1984)Zentralbl. Bakteriol. Para- sitenkd. Infektionska. Hyg. Abt. 1 Orig. Reihe A 257, 68-72
4 Sakaguchi G., Sakaguchi S., Kozaki S. & Taka- hashi M. (1986) FEMS Microbiol. Lett. 33, 23-29
5 Schmidt J. J., Sathyamoorthy V. & DasGupta B. R. (1984) Biochem. Biophys. Res. Comrnun. 119,900-904
6 Schmidt J. J., Sathyamoorthy V. & DasGupta B.R. (1985) ,Arch. Biochem. Biophys. 238, 544-548
7 DasGupta B .R . & Sathyamoorthy V. (1984) Toxicon 22,415-424
8 Sugii S. & Sakaguchi G. (1975) Infect. Immun. 12, 1262-1270
9 Kozaki S., Sakaguchi S. & Sakaguchi G. (197?) Infect. lmmun. 10,750-756
10 DasGupta B. R. & Sugiyama H. (1976) Infect. Immun. 14, 680-686
11 DasGupta B. R. & Woody M. A. (1984) Toxicon 22, 312-315
12 DasGupta B. R. & Sugiyama H. (1977) in: Per- spectives in Toxinology (Bernheimer A .W. , ed.), John Wiley & Sons, New York, pp. 87-119
13 Simpson L. L. (1986) Annu. Rev. Pharmacol. Toxicol. 26, 427-453
14 Eisel U., Jarausch W., Goretzki K., Henschen A., Engels J., Weller U., Hudel M., Habermann E. & Heiner N. (1986) EMBO J. 5, 2495-2502
15 Weber K. & Osburn M. (1969) J. Biol. Chem. 244, 4406-4412
16 DasGupta B. R. & Sugiyama H. (1972) Bio- chem. Biophys. Res. Commun. 48, 108-112
17 Hunkapiller M. W., Lujan E., Ostrander F. & Hood L .E . (1983) Methods Enzymoi. 91, 227-236
18 Esch F. S. (1984) Anal. Biochem. 136, 39-47 19 Heinrickson R, L. & Meredith S.C. (1984)
Anal. Biochem. 136, 65-74 20 Dast3upta B. R. & Rasmussen S. (1984) Arch.
Biochem. Biophys. 232, 172-178 21 Das(3upta B. R. & Rasmussen S. (1981) Bio-
chem. Biophys. Res. Commun. 101, 1209-1215 22 Sathyamoorthy V. & DasGupta B. R. (1985) J.
Biol. Chem. 260, 10461-10466 23 DasGupta B. R. & Sugiyama H. (1977) Can. J.
Microbiol. 23, 1257-1260 24 Shone C. C., Hambleton P. & Melling J. (1985)
Eur. J. Biochem. 151, 75-82 25 Schmitt A., Dreyer F. & John C. (1981)Naunyn-
Schmiedeberg's Arch. Pharmacol. 317, 326-330 26 Simpson L. L. (1984)J. Pharmacol. Exp. Ther.
229, 182-187 27 Bandyopadhyay S., Clark A.W. , DasGupta
B. R. & Sathyamoorthy V. (1987)J. Biol. Chem. 262, 2660-2663
