- Email: [email protected]
A 45-kDa acetylcholinesterase protoxin of Aeromonas hydrophila: purification and immunogenicity in fish
FEMS Microbiology Letters 211 (2002) 23^27
www.fems-microbiology.org
A 45-kDa acetylcholinesterase protoxin of Aeromonas hydrophila: puri¢cation and immunogenicity in ¢sh M.J. Pe¤rez a , L.A. Rodr|¤guez a , A. Ferna¤ndez-Briera b , T.P. Nieto a
a;
Departamento de Biolog|¤a Funcional y Ciencias de la Salud, Universidad de Vigo, Facultad de Ciencias, E-36200 Vigo, Spain b Departamento de Bioqu|¤mica, Gene¤tica e Inmunolog|¤a, Universidad de Vigo, Facultad de Ciencias, E-36200 Vigo, Spain Received 6 December 2001; received in revised form 11 March 2002; accepted 16 March 2002 First published online 12 April 2002
Abstract A rabbit antiserum to the 15-kDa acetylcholinesterase toxin neutralised the lethal effect of the 15-kDa toxin of Aeromonas hydrophila when injected into trout. However, immunisation of fish with the 15-kDa toxoid failed to induce an antibody response, and a higher molecular mass form of this toxin was purified from the extracellular products with the aim of inducing an immune response in fish. The optimal conditions for production of extracellular products by A. hydrophila strain B32 were studied to increase the concentration of this protoxin. The extracellular products were fractionated by molecular exclusion chromatography to yield a purified protoxin with an estimated molecular mass of 45 kDa by SDS^PAGE and which gave a positive reaction in Western blotting with the rabbit anti-15-kDa toxin serum. Since the 45-kDa protoxin showed lower specific acetylcholinesterase activity than the active 15-kDa toxin, the behaviour of the active site was studied using specific inhibitors. This 45-kDa protoxin was 13.3-fold less toxic than the 15-kDa toxin and induced antibody production in fish. : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Acetylcholinesterase protoxin; Puri¢cation ; Ichthyotoxin; Aeromonas hydrophila
logical response in rainbow trout were studied to evaluate its potential use as a vaccine antigen for ¢sh.
1. Introduction In a previous study the puri¢cation from Aeromonas hydrophila cultures of a lethal ichthyotoxin with acetylcholinesterase activity was reported [1]. This toxin was described as an extracellular component of molecular mass 96 kDa that was cleaved into an active form of 15 kDa by other components of the crude extracellular products [2]. It was 300 times more toxic than crude culture supernatant with a lethal dose in ¢sh of 50 ng g31 [1]. It had no proteolytic, haemolytic or cytolytic activity [1]. This toxin is widely distributed among motile Aeromonas strains isolated from clinical, ¢sh and environmental sources [2,3] and also amongst Vibrio strains [4]. The current paper deals with the puri¢cation and characterisation of a high molecular mass intermediate form of this toxin. The 50% lethal dose (LD50 ) and the immuno-
* Corresponding author. Tel. : +34 (986) 81 23 98; Fax : +34 (986) 81 25 56. E-mail address : [email protected] (T.P. Nieto).
2. Materials and methods A. hydrophila strain B32 [1] was passaged three times through ¢sh according to Ferna¤ndez et al. [5] before use. The extracellular products were obtained using the cellophane overlay technique [6]. Tryptone soy agar (TSA) plates were inoculated with 2.5 ml of an overnight culture. After incubation at 20‡C for 16 or 24 h, the cellophane overlay cultures were harvested in sterile phosphate bu¡ered saline (PBS), pH 7.4, at room temperature or 4‡C, with or without metallo- and serine-protease inhibitors (50 mM EDTA, Merck, and 10 mM phenylmethanesulfonyl £uoride (PMSF), Sigma, respectively), centrifuged at 9000 rpm at 4‡C for 20 min and the supernatants sterilised by 0.22-Wm membrane ¢ltration (Corning). Crude culture supernatant was concentrated 10.6-fold in a vacuum concentrator (Savant), and 3.9 ml of crude culture supernatant (about 21 mg protein) were applied in 100-Wl volumes to a TSK-G3000SW molecular exclusion column (Perkin-
0378-1097 / 02 / $22.00 : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII : S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 0 6 3 3 - X
FEMSLE 10453 23-5-02
24
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
Elmer) equilibrated with 0.05 M K2 HPO4 (adjusted to pH 7.7 with HCl). The sample was eluted in this bu¡er at a £ow rate of 1 ml min31 and the column was calibrated using albumin (Sigma) 67 kDa and carbonic anhydrase (Sigma) 30 kDa. Peaks were collected in separate fractions, the procedure was repeated several times, and homologous peaks were pooled. Fractions obtained were concentrated in a Speedvac (Savant) to a volume of 2.5 ml, desalted using Sephadex G-25 columns (Pharmacia) resulting in 3.5 ml ¢nal volume of each fraction, and frozen at 380‡C until use. Each eluted peak and crude culture supernatant were analysed for protein concentration (Bradford method [7]), acetylcholinesterase, proteolytic and haemolytic activities, and detection of toxin by Western blotting (see below). Acetylcholinesterase activity in eluted peaks and crude culture supernatant was determined using a commercial kit (Cholinesterase No. 420, colorimetric; Sigma) following the manufacturer’s instructions. One Rappaport unit is that amount of cholinesterase which will liberate 1 Wmole of acetic acid from acetylcholine in 30 min at 25‡C at pH 7.8 [8]. Acetylcholinesterase activity of the protoxin was also assayed by the method of Ellman et al. [9]. One unit of acetylcholinesterase activity was de¢ned as the amount of enzyme which hydrolysed 1 Wmole of acetylthiocholine per hour at 37‡C under the assay conditions. Proteolytic (caseinase) activity assays were performed as previously described [10]. One unit of activity was de¢ned as a change of 0.01 in OD280 after 15 min incubation with substrate at 22‡C. Haemolytic activity was measured using rainbow trout erythrocytes in microtitre plates after incubation for 2.5 h at 22‡C, as previously described [11]. One unit of haemolytic activity was de¢ned as the reciprocal of the highest dilution of the sample producing complete haemolysis. Protein purity and molecular mass of HPLC eluted peaks were estimated by sodium dodecyl sulfate^12% polyacrylamide gel electrophoresis (SDS^PAGE) according to Laemmli [12], stained with silver stain [13] or transferred to nitrocellulose membranes, with immunostaining performed according to Rodriguez et al. [2] using rabbit anti-15-kDa toxin serum previously obtained by Nieto et al. [1]. Kinetic parameters, Km and Vmax , were determined with acetylthiocholine as substrate (0.05^4 mM concentration range) [9]. The reaction mixture was as described above. The kinetic parameters were calculated using the non-linear regression data analysis program ‘Enz¢tter’ (Elsiever, Biosoft). Inhibition studies were performed using 1033 M eserine, 3U1033 M PMSF or 1035 M 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide (all inhibitors supplied by Sigma). The LD50 was determined and the seroneutralisation experiments were done using rainbow trout (2 g and 20 g, respectively) acclimatised to aquarium conditions at
15‡C for 10 days. Serially diluted doses of puri¢ed 45kDa protoxin and a mixture of rabbit anti-15-kDa toxin serum with puri¢ed 15-kDa toxin or crude culture supernatant were injected intraperitoneally (12 ¢sh per dose). Mortalities recorded over 72 h were used to calculate the LD50 by the Reed and Muench method [14]. Controls were injected intraperitoneally with sterile PBS or serum. Rabbit anti-15-kDa toxin serum was obtained according to Rodr|¤guez et al. [2]. To obtain trout anti-45-kDa protoxin and anti-15-kDa toxin sera, each pure component (the 15-kDa toxin was puri¢ed as previously described [1]) was treated by adding formalin (40% w/v) to a concentration of 3% (v/v), overnight incubation at room temperature and dialysis against PBS at 4‡C for 12 h. Formalised 45-kDa protoxin and 15kDa toxin were used as toxoid and injected intraperitoneally into 300-g rainbow trout (six ¢sh) at a dose of 45 Wg protein/¢sh. Twenty days after immunisation a similar booster injection was given and 29 days later sera were collected, pooled and stored at 320‡C until use [15]. As a control, the same protocol was performed with sterile PBS mixed with Freund’s complete adjuvant. SDS^PAGE [12] was used to separate and transfer the 45-kDa protoxin, 15-kDa toxin and crude culture supernatant to nitrocellulose membranes. Nitrocellulose was incubated with undiluted rainbow trout anti-45-kDa protoxin or anti-15-kDa toxin serum for 3 h; the second antibody solution was rabbit anti-rainbow trout IgM diluted 1/250 in antibody bu¡er incubated for 1 h, and the last antibody solution was goat anti-rabbit Ig conjugated to alkaline phosphatase (Bio-Rad) diluted 1/3000 incubated for 1 h. The immunostaining was performed according to Rodr|¤guez et al. [2].
3. Results and discussion In previous papers a protein of 15 kDa with acetylcholinesterase activity was described as the toxin for rainbow trout [1] and was used as an immunogen in a rabbit to produce antiserum [2]. Further experiments reported here show that the lethal e¡ect of the 15-kDa toxin in trout could be neutralised by the rabbit anti-15-kDa toxin antiserum (Table 1). The lethality of crude culture supernatant was also partially neutralised, con¢rming that the 15-kDa toxin was a major, but not the sole, toxic component of culture supernatant as previously described [1]. Because in preliminary experiments we were unable to induce an antibody response in rainbow trout using the 15-kDa formalised toxin, we now investigated whether a formalised protoxin, with a higher molecular mass, would be immunogenic in the trout. Previous experiments had shown that the lethal 15-kDa toxin was secreted as a high molecular mass component that was cleaved at room temperature [2]. Experiments conducted to detect this high molecular mass protoxin in
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
25
Table 1 Neutralisation of 15-kDa toxin and crude culture supernatant using rabbit anti-15-kDa toxin serum Samplea
ng g31 ¢sh
Serum volume (Wl)
No. dead ¢sh/No. injected ¢she
Toxinb Toxin+serumc Toxin+serum Crude culture supernatantd Crude culture supernatant Crude culture supernatant+serum Crude culture supernatant+serum
50 50 50 18 000 15 000 15 000 15 000
^ 50 70 ^ ^ 200 400
5/12 5/12 0/12 10/12 5/12 5/12 2/12
a
The sample was prepared by blending the stated quantity of nanograms of protein (toxin or culture supernatant with the rabbit anti-15-kDa toxin serum volume before injection into ¢sh. b Pure 15-kDa toxin with an LD50 of 50 ng g31 ¢sh (3-g trout) [1]. c Rabbit anti-15-kDa toxin antiserum [2]. d Culture supernatant obtained after 48 h incubation at 25‡C, with LD50 15 Wg g31 ¢sh (3-g trout) [1]. e Twelve rainbow trout (20 g) per experiment were injected. Fish injected with serum were used as control. Mortalities were monitored over 72 h.
culture supernatant are shown in Table 2. When the culture supernatant was collected after 16 h incubation in PBS at 4‡C, the acetylcholinesterase activity was 3.32 RU mg31 and the rabbit anti-15-kDa toxin antiserum detected a band of 45 kDa by Western blotting. Further experiments, summarised in Table 2, suggested that a serine protease in the crude culture supernatant activated the 45-kDa protoxin at room temperature while a metalloprotease broke the 45-kDa protoxin to smaller molecular mass fragments of 15 kDa ; the latter was not present in culture supernatant obtained after 16 h incubation. On the basis of these results, TSA plates of A. hydrophila strain B32 were incubated for 16 h at 20‡C and culture supernatant was harvested with sterile PBS at 4‡C. This culture supernatant showed proteolytic and haemolytic activities (Table 3). A total of 20.67 mg culture supernatant protein (as 0.1-ml consecutive samples) were applied to a column for molecular exclusion chromatography. The elution pro¢le showed seven peaks (Table 3). Peaks 3, 4, and 6 showing acetylcholinesterase activity were not considered for puri¢cation because of the presence of other components with haemolytic and/or proteolytic activities. Only peak 7 had no detectable protease or haemolytic activity. This fraction with acetylcholinesterase activity showed a retention time of 18 min with an estimated molecular mass less than 30 kDa. This peak gave a
single band of 45 kDa in SDS^PAGE and an immunological reaction in Western blotting with rabbit anti-15-kDa toxin serum. This protoxin had low speci¢c Rappaport acetylcholinesterase activity (6.69 RU mg31 ) compared to the previously puri¢ed 15-kDa toxin (31 650 RU mg31 ) [1]. The puri¢cation ratio of this 45-kDa protoxin based on speci¢c RU data was only 2.07, but this is probably because this component is an intermediate toxin form and activation is a complex process. Previous results demonstrated the existence of intermediate forms recognised by Western blotting with molecular masses of 94, 60, 45 and 30 kDa [2,4], with low acetylcholinesterase activity, but it was impossible to purify the highest molecular mass native form. The recovery of the 45-kDa protoxin (with 24.63% of the total RU of culture supernatant) is similar to the values obtained for the puri¢cation of Lhaemolysin in one step [16] and higher than that obtained for that toxin in two steps [17] and for the cytotonic enterotoxin [18]. Some characteristics of the acetylcholinesterase activity of the 45-kDa protoxin were determined. The e¡ect of acetylthiocholine substrate concentrations (ranging from 0.1 to 4 mM) and the kinetic parameters Km (Michaelis^ Menten constant) and Vmax (maximal velocity), using Ellman’s method with 50 Wg of protoxin were determined. Activity at low substrate concentration (0.01^0.08 mM)
Table 2 Assays performed to determine the optimal conditions for obtaining the protoxin from ECP Samplea
Incubation time (h)
Temperature (‡C) for collecting ECP
RU mg31
Western blot
MM (kDa)
ECP in PBS ECP in PBS ECP in PBS ECP+EDTA ECP+EDTA ECP+PMSF ECP+PMSF
24 24 16 24 16 24 24
RT 4 4 RT RT RT 4
19.60 1.1 3.32 25 23.28 1.1 1.1
+ + + + + + +
15 15 45 45 45 15 15
a
ECP collected with or without metallo- and serine-protease inhibitors (50 mM EDTA and 10 mM PMSF, respectively). ECP, culture supernatant ; MM, molecular mass of bands detected in Western blotting ; RU mg31 , Rappaport units of acetylcholinesterase activity per mg of protein; RT, room temperature (20‡C); +, positive reaction in Western blot using rabbit anti-15-kDa toxin antiserum.
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
26
Table 3 Acetylcholinesterase, proteolytic and haemolytic activities and detection of protoxin by Western blot of fractions obtained from molecular exclusion chromatography Fraction RT
Total vol. (ml) Total protein (mg) RU mg31
Total RU
Total PA
PA mg31
Total HU
HU mg31
Western blotting (MM)
ECP 1 2 3 4 5 6 7
3.9 3.5 3.5 3.5 3.5 3.5 3.5 3.5
66.58 NT 0 7.56 51.1 0 25.76 16.40
21 083 NT NT 364 6 062 2 303 742 0
1 020 NT NT 70.40 5 885.43 796.89 94.64 0
13 229 NT NT 1 120 140 140 0 0
640 NT NT 216.63 135.92 48.44 0 0
+(96, 67, 45, 30, 15) NT NT +(67, 45, 30) +(major 30) NT +(45) +(45)
1 7 10 12 14 16 18
20.67 0.20 (1%) 0.41 (2%) 5.17 (25%) 1.03 (5%) 2.89 (14%) 7.84 (38%) 2.45 (12%)
3.22 NT 0 1.46 49.61 0 3.29 6.69
Recovered total protein: 19.99 mg (96.71%). ECP, crude culture supernatant; RT, retention time in minutes; the standards used to calibrated the column were albumin (67 kDa) with a RT of 8.5 min and carbonic anhydrase (30 kDa) with RT of 10.12 min ; NT, not tested ; RU mg31 , Rappaport units per mg protein ; PA mg31 , proteolytic units per mg protein; HU mg31 , haemolytic units per mg protein; +, positive reaction in Western blotting using rabbit anti-15-kDa toxin antiserum ; MM, molecular mass of bands in SDS^PAGE also detected by Western blotting.
was undetectable because the protoxin had a very low activity. On the other hand, assays at substrate concentrations greater than 5 mM showed spontaneous activity due to non-enzymatic hydrolysis of the substrate or substrate inhibition. The Michaelis^Menten representation of acetylcholinesterase activity of the protoxin showed a hyperbolic curve and representation of Eadie^Hofstee con¢rmed the hyperbolic kinetics. Kinetic parameters calculated by the non-linear regression program Enz¢tter were 0.64 mM Km and 3.16 U mg31 Vmax . This value of Km was higher than that obtained in similar conditions for rat brain acetylcholinesterase (0.055 mM) [19] and invertebrate acetylcholinesterases (0.03^0.08 mM) [20,21], and close to the Km determined for ¢sh brain acetylcholinesterase (0.25 mM) [19]. The residual acetylcholinesterase activity of the 45-kDa protoxin in the presence of inhibitors is shown in Table 4. Eserine (1 mM) inhibited totally the acetylcholinesterase activity, as is typical of acetylcholinesterases [19]. However, 3 mM PMSF or 10 WM 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide produced inhibition of 75% and 50% of the acetylcholinesterase activity, respectively. The PMSF e¡ect suggested that the 45-kDa protoxin is similar to mammalian acetylcholinesterase and di¡erent from ¢sh acetylcholinesterase [19]. The partial inhibition of the 45-kDa protoxin by 1035 M 1,5-bis(4allyldimethylammoniumphenyl)pentan-3-one dibromide showed di¡erences from the typical acetylcholinesterases described [22]. The LD50 for this 45-kDa protoxin was 2 Wg g31 rainTable 4 Inhibition of acetylcholinesterase activity using di¡erent inhibitors Inhibitor
[Inhibitor]
Residual activity (%)
None Eserine PMSF BW
^ 1 mM 3 mM 10 WM
100 0 26.06 51.38
BW, 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide.
bow trout with no gross external and/or internal lesions observable in the dead ¢sh. A decrease in the time of mortality was detected with an increase of injected protoxin dose. The protoxin was 13.3-fold less toxic than the pure active toxin [1]. A positive reaction in Western blotting using undiluted trout anti-45-kDa protoxin serum was detected against the 45-kDa protoxin, 15-kDa toxin and crude ECP. No reaction was found using sera from trout immunised with the 15-kDa toxoid. In order to consider this 45-kDa protoxin as an antigen to be used in ¢sh vaccination strategies against motile Aeromonas infection, experiments on seroneutralisation and immunostimulation should be conducted to determine if the 45-kDa protoxin induces adequate protection in ¢sh.
Acknowledgements This work was supported by grants from the Universidad de Vigo (1993-1994), DGICYT (PB89-0547), DGICYT (MAR97-1188-C02-01) and Xunta de Galicia (PGIDT00PX130111PR). M.J.P. acknowledges to Conseller|¤a de Educacio¤n e Ordenacio¤n Universitaria de la Xunta de Galicia for a research fellowship.
References [1] Nieto, T.P., Santos, Y., Rodr|¤guez, L.A. and Ellis, A.E. (1991) An extracellular acetylcholinesterase produced by Aeromonas hydrophila is a major lethal toxin for ¢sh. Microb. Pathogen. 11, 101^110. [2] Rodr|¤guez, L.A., Ferna¤ndez, A.I.G. and Nieto, T.P. (1993) Production of the lethal acetylcholinesterase toxin by di¡erent Aeromonas hydrophila strains. J. Fish Dis. 16, 73^78. [3] Pe¤rez, M.J., Rodr|¤guez, L.A., Ferna¤ndez, A.I.G. and Nieto, T.P. (1995) Distribucio¤n de la ictiotoxina letal de Aeromonas hydrophila entre cepas de diferentes or|¤genes. In: Actas V Congr. Nac. Acuicult. (Castello i Orvay, F. and Calderer i Reig, A., Eds.), pp. 760^764 Universidad de Barcelona, Barcelona. [4] Pe¤rez, M.J., Rodr|¤guez, L.A. and Nieto, T.P. (1998) The acetylcho-
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
[5]
[6] [7]
[8]
[9]
[10]
[11]
[12] [13]
linesterase ichthyotoxin is a common component in the extracellular products of Vibrionaceae strains. J. Appl. Microbiol. 84, 47^52. Ferna¤ndez, A.I.G., Pe¤rez, M.J., Rodr|¤guez, L.A. and Nieto, T.P. (1995) Surface phenotypic characteristics and virulence of Spanish isolates of Aeromonas salmonicida after passage though ¢sh. Appl. Environ. Microbiol. 61, 20110^20112. Liu, P.V. (1957) Survey of hemolysin production among species of pseudomonads. J. Bacteriol. 74, 718^727. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of micrograms of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248^254. Rappaport, F., Fischl, J. and Pinto, N. (1959) An improved method for the estimation of cholinesterase activity in serum. Clin. Chim. Acta 4, 227. Ellman, G.L., Courtney, D., Andres, V. and Featherstone, R.M. (1961) A new rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88^95. Nieto, T.P. and Ellis, A.E. (1986) Characterization of extracellular metallo- and serine-proteases of Aeromonas hydrophila strain B51 . J. Gen. Microbiol. 132, 1975^1979. Hastings, T.S. and Ellis, A.E. (1985) Di¡erences in the production of haemolytic and proteolytic activities of various isolates of Aeromonas salmonicida. In: Fish and Shell¢sh Pathology (Ellis, A.E., Ed.), pp. 98^106. Academic Press, London. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680^685. Morrisey, J.M. (1981) Silver stain for proteins in polyacrylamide gels: a modi¢ed procedure with enhanced uniform sensitivity. Anal. Biochem. 117, 307^310.
27
[14] Reed, L.J. and Muench, H. (1938) A simple method of estimating ¢fty percent endpoints. Am. J. Hyg. 27, 493^497. [15] Hastings, T.S. and Ellis, A.E. (1988) The humoral response of rainbow trout Salmo gairdneri Richardson, and rabbits to Aeromonas salmonicida extracellular products. J. Fish Dis. 11, 147^160. [16] Loychern, A., Charoensini, K. and Ratanabangkoon, K. (1994) A simple puri¢cation of Aeromonas hydrophila hemolysin. J. Nat. Toxins 3, 1^6. [17] Buckley, J.T., Halasa, L.N., Lund, K.D. and MacIntyre, S. (1981) Puri¢cation and some properties of the hemolytic toxin aerolysin. Can. J. Biochem. 59, 430^435. [18] Gosling, P.J., Turnbull, P.C.B., Lighfoot, N.F., Pether, J.V.S. and Lewis, R.J. (1993) Isolation and puri¢cation of Aeromonas sobria cytotonic enterotoxin and L-haemolysin. J. Med. Microbiol. 38, 227^234. [19] Moss, D.E. and Farhney, D. (1978) Kinetic analysis of di¡erences in brain acetylcholinesterase from ¢sh or mammalian sources. Biochem. Pharmacol. 27, 2693^2698. [20] Principiato, G.B., Rosi, G., Bocchini, V. and Giovannini, E. (1984) Propionylcholinesterase from Helix pomatia and acetylcholinesterase from Asterias bispinosa: a kinetic comparative study. Comp. Biochem. Physiol. 77B, 211^219. [21] Talesa, V., Contenti, S., Mangiabene, C., Pascolini, R., Rosi, G. and Principiato, G.B. (1990) Propionylcholinesterase from Murex brandaris : comparison with other invertebrate cholinesterases. Comp. Biochem. Physiol. 96C, 39^43. [22] Tornel, P.L., Saez-Valero, J. and Vidal, C.J. (1992) Ricinus communis agglutinin I reacting and non-reacting butyrylcholinesterase in human cerebrospinal £uid. Neurosci. Lett. 145, 98^106.
FEMSLE 10453 23-5-02
www.fems-microbiology.org
A 45-kDa acetylcholinesterase protoxin of Aeromonas hydrophila: puri¢cation and immunogenicity in ¢sh M.J. Pe¤rez a , L.A. Rodr|¤guez a , A. Ferna¤ndez-Briera b , T.P. Nieto a
a;
Departamento de Biolog|¤a Funcional y Ciencias de la Salud, Universidad de Vigo, Facultad de Ciencias, E-36200 Vigo, Spain b Departamento de Bioqu|¤mica, Gene¤tica e Inmunolog|¤a, Universidad de Vigo, Facultad de Ciencias, E-36200 Vigo, Spain Received 6 December 2001; received in revised form 11 March 2002; accepted 16 March 2002 First published online 12 April 2002
Abstract A rabbit antiserum to the 15-kDa acetylcholinesterase toxin neutralised the lethal effect of the 15-kDa toxin of Aeromonas hydrophila when injected into trout. However, immunisation of fish with the 15-kDa toxoid failed to induce an antibody response, and a higher molecular mass form of this toxin was purified from the extracellular products with the aim of inducing an immune response in fish. The optimal conditions for production of extracellular products by A. hydrophila strain B32 were studied to increase the concentration of this protoxin. The extracellular products were fractionated by molecular exclusion chromatography to yield a purified protoxin with an estimated molecular mass of 45 kDa by SDS^PAGE and which gave a positive reaction in Western blotting with the rabbit anti-15-kDa toxin serum. Since the 45-kDa protoxin showed lower specific acetylcholinesterase activity than the active 15-kDa toxin, the behaviour of the active site was studied using specific inhibitors. This 45-kDa protoxin was 13.3-fold less toxic than the 15-kDa toxin and induced antibody production in fish. : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Acetylcholinesterase protoxin; Puri¢cation ; Ichthyotoxin; Aeromonas hydrophila
logical response in rainbow trout were studied to evaluate its potential use as a vaccine antigen for ¢sh.
1. Introduction In a previous study the puri¢cation from Aeromonas hydrophila cultures of a lethal ichthyotoxin with acetylcholinesterase activity was reported [1]. This toxin was described as an extracellular component of molecular mass 96 kDa that was cleaved into an active form of 15 kDa by other components of the crude extracellular products [2]. It was 300 times more toxic than crude culture supernatant with a lethal dose in ¢sh of 50 ng g31 [1]. It had no proteolytic, haemolytic or cytolytic activity [1]. This toxin is widely distributed among motile Aeromonas strains isolated from clinical, ¢sh and environmental sources [2,3] and also amongst Vibrio strains [4]. The current paper deals with the puri¢cation and characterisation of a high molecular mass intermediate form of this toxin. The 50% lethal dose (LD50 ) and the immuno-
* Corresponding author. Tel. : +34 (986) 81 23 98; Fax : +34 (986) 81 25 56. E-mail address : [email protected] (T.P. Nieto).
2. Materials and methods A. hydrophila strain B32 [1] was passaged three times through ¢sh according to Ferna¤ndez et al. [5] before use. The extracellular products were obtained using the cellophane overlay technique [6]. Tryptone soy agar (TSA) plates were inoculated with 2.5 ml of an overnight culture. After incubation at 20‡C for 16 or 24 h, the cellophane overlay cultures were harvested in sterile phosphate bu¡ered saline (PBS), pH 7.4, at room temperature or 4‡C, with or without metallo- and serine-protease inhibitors (50 mM EDTA, Merck, and 10 mM phenylmethanesulfonyl £uoride (PMSF), Sigma, respectively), centrifuged at 9000 rpm at 4‡C for 20 min and the supernatants sterilised by 0.22-Wm membrane ¢ltration (Corning). Crude culture supernatant was concentrated 10.6-fold in a vacuum concentrator (Savant), and 3.9 ml of crude culture supernatant (about 21 mg protein) were applied in 100-Wl volumes to a TSK-G3000SW molecular exclusion column (Perkin-
0378-1097 / 02 / $22.00 : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII : S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 0 6 3 3 - X
FEMSLE 10453 23-5-02
24
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
Elmer) equilibrated with 0.05 M K2 HPO4 (adjusted to pH 7.7 with HCl). The sample was eluted in this bu¡er at a £ow rate of 1 ml min31 and the column was calibrated using albumin (Sigma) 67 kDa and carbonic anhydrase (Sigma) 30 kDa. Peaks were collected in separate fractions, the procedure was repeated several times, and homologous peaks were pooled. Fractions obtained were concentrated in a Speedvac (Savant) to a volume of 2.5 ml, desalted using Sephadex G-25 columns (Pharmacia) resulting in 3.5 ml ¢nal volume of each fraction, and frozen at 380‡C until use. Each eluted peak and crude culture supernatant were analysed for protein concentration (Bradford method [7]), acetylcholinesterase, proteolytic and haemolytic activities, and detection of toxin by Western blotting (see below). Acetylcholinesterase activity in eluted peaks and crude culture supernatant was determined using a commercial kit (Cholinesterase No. 420, colorimetric; Sigma) following the manufacturer’s instructions. One Rappaport unit is that amount of cholinesterase which will liberate 1 Wmole of acetic acid from acetylcholine in 30 min at 25‡C at pH 7.8 [8]. Acetylcholinesterase activity of the protoxin was also assayed by the method of Ellman et al. [9]. One unit of acetylcholinesterase activity was de¢ned as the amount of enzyme which hydrolysed 1 Wmole of acetylthiocholine per hour at 37‡C under the assay conditions. Proteolytic (caseinase) activity assays were performed as previously described [10]. One unit of activity was de¢ned as a change of 0.01 in OD280 after 15 min incubation with substrate at 22‡C. Haemolytic activity was measured using rainbow trout erythrocytes in microtitre plates after incubation for 2.5 h at 22‡C, as previously described [11]. One unit of haemolytic activity was de¢ned as the reciprocal of the highest dilution of the sample producing complete haemolysis. Protein purity and molecular mass of HPLC eluted peaks were estimated by sodium dodecyl sulfate^12% polyacrylamide gel electrophoresis (SDS^PAGE) according to Laemmli [12], stained with silver stain [13] or transferred to nitrocellulose membranes, with immunostaining performed according to Rodriguez et al. [2] using rabbit anti-15-kDa toxin serum previously obtained by Nieto et al. [1]. Kinetic parameters, Km and Vmax , were determined with acetylthiocholine as substrate (0.05^4 mM concentration range) [9]. The reaction mixture was as described above. The kinetic parameters were calculated using the non-linear regression data analysis program ‘Enz¢tter’ (Elsiever, Biosoft). Inhibition studies were performed using 1033 M eserine, 3U1033 M PMSF or 1035 M 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide (all inhibitors supplied by Sigma). The LD50 was determined and the seroneutralisation experiments were done using rainbow trout (2 g and 20 g, respectively) acclimatised to aquarium conditions at
15‡C for 10 days. Serially diluted doses of puri¢ed 45kDa protoxin and a mixture of rabbit anti-15-kDa toxin serum with puri¢ed 15-kDa toxin or crude culture supernatant were injected intraperitoneally (12 ¢sh per dose). Mortalities recorded over 72 h were used to calculate the LD50 by the Reed and Muench method [14]. Controls were injected intraperitoneally with sterile PBS or serum. Rabbit anti-15-kDa toxin serum was obtained according to Rodr|¤guez et al. [2]. To obtain trout anti-45-kDa protoxin and anti-15-kDa toxin sera, each pure component (the 15-kDa toxin was puri¢ed as previously described [1]) was treated by adding formalin (40% w/v) to a concentration of 3% (v/v), overnight incubation at room temperature and dialysis against PBS at 4‡C for 12 h. Formalised 45-kDa protoxin and 15kDa toxin were used as toxoid and injected intraperitoneally into 300-g rainbow trout (six ¢sh) at a dose of 45 Wg protein/¢sh. Twenty days after immunisation a similar booster injection was given and 29 days later sera were collected, pooled and stored at 320‡C until use [15]. As a control, the same protocol was performed with sterile PBS mixed with Freund’s complete adjuvant. SDS^PAGE [12] was used to separate and transfer the 45-kDa protoxin, 15-kDa toxin and crude culture supernatant to nitrocellulose membranes. Nitrocellulose was incubated with undiluted rainbow trout anti-45-kDa protoxin or anti-15-kDa toxin serum for 3 h; the second antibody solution was rabbit anti-rainbow trout IgM diluted 1/250 in antibody bu¡er incubated for 1 h, and the last antibody solution was goat anti-rabbit Ig conjugated to alkaline phosphatase (Bio-Rad) diluted 1/3000 incubated for 1 h. The immunostaining was performed according to Rodr|¤guez et al. [2].
3. Results and discussion In previous papers a protein of 15 kDa with acetylcholinesterase activity was described as the toxin for rainbow trout [1] and was used as an immunogen in a rabbit to produce antiserum [2]. Further experiments reported here show that the lethal e¡ect of the 15-kDa toxin in trout could be neutralised by the rabbit anti-15-kDa toxin antiserum (Table 1). The lethality of crude culture supernatant was also partially neutralised, con¢rming that the 15-kDa toxin was a major, but not the sole, toxic component of culture supernatant as previously described [1]. Because in preliminary experiments we were unable to induce an antibody response in rainbow trout using the 15-kDa formalised toxin, we now investigated whether a formalised protoxin, with a higher molecular mass, would be immunogenic in the trout. Previous experiments had shown that the lethal 15-kDa toxin was secreted as a high molecular mass component that was cleaved at room temperature [2]. Experiments conducted to detect this high molecular mass protoxin in
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
25
Table 1 Neutralisation of 15-kDa toxin and crude culture supernatant using rabbit anti-15-kDa toxin serum Samplea
ng g31 ¢sh
Serum volume (Wl)
No. dead ¢sh/No. injected ¢she
Toxinb Toxin+serumc Toxin+serum Crude culture supernatantd Crude culture supernatant Crude culture supernatant+serum Crude culture supernatant+serum
50 50 50 18 000 15 000 15 000 15 000
^ 50 70 ^ ^ 200 400
5/12 5/12 0/12 10/12 5/12 5/12 2/12
a
The sample was prepared by blending the stated quantity of nanograms of protein (toxin or culture supernatant with the rabbit anti-15-kDa toxin serum volume before injection into ¢sh. b Pure 15-kDa toxin with an LD50 of 50 ng g31 ¢sh (3-g trout) [1]. c Rabbit anti-15-kDa toxin antiserum [2]. d Culture supernatant obtained after 48 h incubation at 25‡C, with LD50 15 Wg g31 ¢sh (3-g trout) [1]. e Twelve rainbow trout (20 g) per experiment were injected. Fish injected with serum were used as control. Mortalities were monitored over 72 h.
culture supernatant are shown in Table 2. When the culture supernatant was collected after 16 h incubation in PBS at 4‡C, the acetylcholinesterase activity was 3.32 RU mg31 and the rabbit anti-15-kDa toxin antiserum detected a band of 45 kDa by Western blotting. Further experiments, summarised in Table 2, suggested that a serine protease in the crude culture supernatant activated the 45-kDa protoxin at room temperature while a metalloprotease broke the 45-kDa protoxin to smaller molecular mass fragments of 15 kDa ; the latter was not present in culture supernatant obtained after 16 h incubation. On the basis of these results, TSA plates of A. hydrophila strain B32 were incubated for 16 h at 20‡C and culture supernatant was harvested with sterile PBS at 4‡C. This culture supernatant showed proteolytic and haemolytic activities (Table 3). A total of 20.67 mg culture supernatant protein (as 0.1-ml consecutive samples) were applied to a column for molecular exclusion chromatography. The elution pro¢le showed seven peaks (Table 3). Peaks 3, 4, and 6 showing acetylcholinesterase activity were not considered for puri¢cation because of the presence of other components with haemolytic and/or proteolytic activities. Only peak 7 had no detectable protease or haemolytic activity. This fraction with acetylcholinesterase activity showed a retention time of 18 min with an estimated molecular mass less than 30 kDa. This peak gave a
single band of 45 kDa in SDS^PAGE and an immunological reaction in Western blotting with rabbit anti-15-kDa toxin serum. This protoxin had low speci¢c Rappaport acetylcholinesterase activity (6.69 RU mg31 ) compared to the previously puri¢ed 15-kDa toxin (31 650 RU mg31 ) [1]. The puri¢cation ratio of this 45-kDa protoxin based on speci¢c RU data was only 2.07, but this is probably because this component is an intermediate toxin form and activation is a complex process. Previous results demonstrated the existence of intermediate forms recognised by Western blotting with molecular masses of 94, 60, 45 and 30 kDa [2,4], with low acetylcholinesterase activity, but it was impossible to purify the highest molecular mass native form. The recovery of the 45-kDa protoxin (with 24.63% of the total RU of culture supernatant) is similar to the values obtained for the puri¢cation of Lhaemolysin in one step [16] and higher than that obtained for that toxin in two steps [17] and for the cytotonic enterotoxin [18]. Some characteristics of the acetylcholinesterase activity of the 45-kDa protoxin were determined. The e¡ect of acetylthiocholine substrate concentrations (ranging from 0.1 to 4 mM) and the kinetic parameters Km (Michaelis^ Menten constant) and Vmax (maximal velocity), using Ellman’s method with 50 Wg of protoxin were determined. Activity at low substrate concentration (0.01^0.08 mM)
Table 2 Assays performed to determine the optimal conditions for obtaining the protoxin from ECP Samplea
Incubation time (h)
Temperature (‡C) for collecting ECP
RU mg31
Western blot
MM (kDa)
ECP in PBS ECP in PBS ECP in PBS ECP+EDTA ECP+EDTA ECP+PMSF ECP+PMSF
24 24 16 24 16 24 24
RT 4 4 RT RT RT 4
19.60 1.1 3.32 25 23.28 1.1 1.1
+ + + + + + +
15 15 45 45 45 15 15
a
ECP collected with or without metallo- and serine-protease inhibitors (50 mM EDTA and 10 mM PMSF, respectively). ECP, culture supernatant ; MM, molecular mass of bands detected in Western blotting ; RU mg31 , Rappaport units of acetylcholinesterase activity per mg of protein; RT, room temperature (20‡C); +, positive reaction in Western blot using rabbit anti-15-kDa toxin antiserum.
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
26
Table 3 Acetylcholinesterase, proteolytic and haemolytic activities and detection of protoxin by Western blot of fractions obtained from molecular exclusion chromatography Fraction RT
Total vol. (ml) Total protein (mg) RU mg31
Total RU
Total PA
PA mg31
Total HU
HU mg31
Western blotting (MM)
ECP 1 2 3 4 5 6 7
3.9 3.5 3.5 3.5 3.5 3.5 3.5 3.5
66.58 NT 0 7.56 51.1 0 25.76 16.40
21 083 NT NT 364 6 062 2 303 742 0
1 020 NT NT 70.40 5 885.43 796.89 94.64 0
13 229 NT NT 1 120 140 140 0 0
640 NT NT 216.63 135.92 48.44 0 0
+(96, 67, 45, 30, 15) NT NT +(67, 45, 30) +(major 30) NT +(45) +(45)
1 7 10 12 14 16 18
20.67 0.20 (1%) 0.41 (2%) 5.17 (25%) 1.03 (5%) 2.89 (14%) 7.84 (38%) 2.45 (12%)
3.22 NT 0 1.46 49.61 0 3.29 6.69
Recovered total protein: 19.99 mg (96.71%). ECP, crude culture supernatant; RT, retention time in minutes; the standards used to calibrated the column were albumin (67 kDa) with a RT of 8.5 min and carbonic anhydrase (30 kDa) with RT of 10.12 min ; NT, not tested ; RU mg31 , Rappaport units per mg protein ; PA mg31 , proteolytic units per mg protein; HU mg31 , haemolytic units per mg protein; +, positive reaction in Western blotting using rabbit anti-15-kDa toxin antiserum ; MM, molecular mass of bands in SDS^PAGE also detected by Western blotting.
was undetectable because the protoxin had a very low activity. On the other hand, assays at substrate concentrations greater than 5 mM showed spontaneous activity due to non-enzymatic hydrolysis of the substrate or substrate inhibition. The Michaelis^Menten representation of acetylcholinesterase activity of the protoxin showed a hyperbolic curve and representation of Eadie^Hofstee con¢rmed the hyperbolic kinetics. Kinetic parameters calculated by the non-linear regression program Enz¢tter were 0.64 mM Km and 3.16 U mg31 Vmax . This value of Km was higher than that obtained in similar conditions for rat brain acetylcholinesterase (0.055 mM) [19] and invertebrate acetylcholinesterases (0.03^0.08 mM) [20,21], and close to the Km determined for ¢sh brain acetylcholinesterase (0.25 mM) [19]. The residual acetylcholinesterase activity of the 45-kDa protoxin in the presence of inhibitors is shown in Table 4. Eserine (1 mM) inhibited totally the acetylcholinesterase activity, as is typical of acetylcholinesterases [19]. However, 3 mM PMSF or 10 WM 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide produced inhibition of 75% and 50% of the acetylcholinesterase activity, respectively. The PMSF e¡ect suggested that the 45-kDa protoxin is similar to mammalian acetylcholinesterase and di¡erent from ¢sh acetylcholinesterase [19]. The partial inhibition of the 45-kDa protoxin by 1035 M 1,5-bis(4allyldimethylammoniumphenyl)pentan-3-one dibromide showed di¡erences from the typical acetylcholinesterases described [22]. The LD50 for this 45-kDa protoxin was 2 Wg g31 rainTable 4 Inhibition of acetylcholinesterase activity using di¡erent inhibitors Inhibitor
[Inhibitor]
Residual activity (%)
None Eserine PMSF BW
^ 1 mM 3 mM 10 WM
100 0 26.06 51.38
BW, 1,5-bis(4-allyldimethylammoniumphenyl)pentan-3-one dibromide.
bow trout with no gross external and/or internal lesions observable in the dead ¢sh. A decrease in the time of mortality was detected with an increase of injected protoxin dose. The protoxin was 13.3-fold less toxic than the pure active toxin [1]. A positive reaction in Western blotting using undiluted trout anti-45-kDa protoxin serum was detected against the 45-kDa protoxin, 15-kDa toxin and crude ECP. No reaction was found using sera from trout immunised with the 15-kDa toxoid. In order to consider this 45-kDa protoxin as an antigen to be used in ¢sh vaccination strategies against motile Aeromonas infection, experiments on seroneutralisation and immunostimulation should be conducted to determine if the 45-kDa protoxin induces adequate protection in ¢sh.
Acknowledgements This work was supported by grants from the Universidad de Vigo (1993-1994), DGICYT (PB89-0547), DGICYT (MAR97-1188-C02-01) and Xunta de Galicia (PGIDT00PX130111PR). M.J.P. acknowledges to Conseller|¤a de Educacio¤n e Ordenacio¤n Universitaria de la Xunta de Galicia for a research fellowship.
References [1] Nieto, T.P., Santos, Y., Rodr|¤guez, L.A. and Ellis, A.E. (1991) An extracellular acetylcholinesterase produced by Aeromonas hydrophila is a major lethal toxin for ¢sh. Microb. Pathogen. 11, 101^110. [2] Rodr|¤guez, L.A., Ferna¤ndez, A.I.G. and Nieto, T.P. (1993) Production of the lethal acetylcholinesterase toxin by di¡erent Aeromonas hydrophila strains. J. Fish Dis. 16, 73^78. [3] Pe¤rez, M.J., Rodr|¤guez, L.A., Ferna¤ndez, A.I.G. and Nieto, T.P. (1995) Distribucio¤n de la ictiotoxina letal de Aeromonas hydrophila entre cepas de diferentes or|¤genes. In: Actas V Congr. Nac. Acuicult. (Castello i Orvay, F. and Calderer i Reig, A., Eds.), pp. 760^764 Universidad de Barcelona, Barcelona. [4] Pe¤rez, M.J., Rodr|¤guez, L.A. and Nieto, T.P. (1998) The acetylcho-
FEMSLE 10453 23-5-02
M.J. Pe¤rez et al. / FEMS Microbiology Letters 211 (2002) 23^27
[5]
[6] [7]
[8]
[9]
[10]
[11]
[12] [13]
linesterase ichthyotoxin is a common component in the extracellular products of Vibrionaceae strains. J. Appl. Microbiol. 84, 47^52. Ferna¤ndez, A.I.G., Pe¤rez, M.J., Rodr|¤guez, L.A. and Nieto, T.P. (1995) Surface phenotypic characteristics and virulence of Spanish isolates of Aeromonas salmonicida after passage though ¢sh. Appl. Environ. Microbiol. 61, 20110^20112. Liu, P.V. (1957) Survey of hemolysin production among species of pseudomonads. J. Bacteriol. 74, 718^727. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of micrograms of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248^254. Rappaport, F., Fischl, J. and Pinto, N. (1959) An improved method for the estimation of cholinesterase activity in serum. Clin. Chim. Acta 4, 227. Ellman, G.L., Courtney, D., Andres, V. and Featherstone, R.M. (1961) A new rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88^95. Nieto, T.P. and Ellis, A.E. (1986) Characterization of extracellular metallo- and serine-proteases of Aeromonas hydrophila strain B51 . J. Gen. Microbiol. 132, 1975^1979. Hastings, T.S. and Ellis, A.E. (1985) Di¡erences in the production of haemolytic and proteolytic activities of various isolates of Aeromonas salmonicida. In: Fish and Shell¢sh Pathology (Ellis, A.E., Ed.), pp. 98^106. Academic Press, London. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680^685. Morrisey, J.M. (1981) Silver stain for proteins in polyacrylamide gels: a modi¢ed procedure with enhanced uniform sensitivity. Anal. Biochem. 117, 307^310.
27
[14] Reed, L.J. and Muench, H. (1938) A simple method of estimating ¢fty percent endpoints. Am. J. Hyg. 27, 493^497. [15] Hastings, T.S. and Ellis, A.E. (1988) The humoral response of rainbow trout Salmo gairdneri Richardson, and rabbits to Aeromonas salmonicida extracellular products. J. Fish Dis. 11, 147^160. [16] Loychern, A., Charoensini, K. and Ratanabangkoon, K. (1994) A simple puri¢cation of Aeromonas hydrophila hemolysin. J. Nat. Toxins 3, 1^6. [17] Buckley, J.T., Halasa, L.N., Lund, K.D. and MacIntyre, S. (1981) Puri¢cation and some properties of the hemolytic toxin aerolysin. Can. J. Biochem. 59, 430^435. [18] Gosling, P.J., Turnbull, P.C.B., Lighfoot, N.F., Pether, J.V.S. and Lewis, R.J. (1993) Isolation and puri¢cation of Aeromonas sobria cytotonic enterotoxin and L-haemolysin. J. Med. Microbiol. 38, 227^234. [19] Moss, D.E. and Farhney, D. (1978) Kinetic analysis of di¡erences in brain acetylcholinesterase from ¢sh or mammalian sources. Biochem. Pharmacol. 27, 2693^2698. [20] Principiato, G.B., Rosi, G., Bocchini, V. and Giovannini, E. (1984) Propionylcholinesterase from Helix pomatia and acetylcholinesterase from Asterias bispinosa: a kinetic comparative study. Comp. Biochem. Physiol. 77B, 211^219. [21] Talesa, V., Contenti, S., Mangiabene, C., Pascolini, R., Rosi, G. and Principiato, G.B. (1990) Propionylcholinesterase from Murex brandaris : comparison with other invertebrate cholinesterases. Comp. Biochem. Physiol. 96C, 39^43. [22] Tornel, P.L., Saez-Valero, J. and Vidal, C.J. (1992) Ricinus communis agglutinin I reacting and non-reacting butyrylcholinesterase in human cerebrospinal £uid. Neurosci. Lett. 145, 98^106.
FEMSLE 10453 23-5-02