- Email: [email protected]
Fasciculin: Modification of carboxyl groups and discussion of structure-activity relationship
0041-0101(95)00155-7
T&con. Vol. 34, No 6, pp. 718-721, 1996 Copyright 8 1996 Elsevier Science Ltd Pnnted in Great Britain. All rights reserved 0041~)101/96 $9.50 + 0.00
SHORT COMMUNICATION FASCICULIN: MODIFICATION OF CARBOXYL GROUPS AND DISCUSSION OF STRUCTURE-ACTIVITY RELATIONSHIP CARLOS CERVERANSKY,’ ROSARIO DURAN’ and EVERT KARLSSON’* ‘Institute de Investigaciones Biologicas Clemente Estable, Avenida ltalia 3318, 11600 Montevideo, Uruguay, and 2Department of Biochemistry, Biomedical Centre, Box 576, 751 23 Uppsala, Sweden. (Received
7 November
1995; Accepted
24 November
1995)
C. Cervenansky, R. Duran and E. Karlsson. Fasciculin: modification of carboxyl groups and discussion of structure-activity relationship. Toxicon 34, 7 18-72 1, 1996.-Norleucine methylester was coupled to carboxylates of fasciculin 2, a snake toxin that inhibits acetylcholinesterase (AChE). This neutralized negative charges but had no effect on the activity, suggesting that carboxyls do not participate in binding to AChE. Earlier results are discussed. Modification of three aromatic amino acids in the peripheral site of AChE, the binding site for fasciculin, decreased the affinity 100 to one million times. Neutralizing the charge of cationic groups of fasciculin lowered the affinity only three to seven times. A change in either the toxin or enzyme part of a binding site should have about the same effect. Since this was not so, it suggests that cationic groups of fasciculin do not bind to aromatic rings in the peripheral site. Copyright 0 1996 Elsevier Science Ltd
Fasciculins are cationic molecules with a very asymmetrical distribution of positive and negative charges (Cerveiiansky ef al., 1991) (Fig. 1). The role of cationic groups has been studied by chemical modification of amino groups and arginine residues (Cerveiiansky et al., 1994, 1995). We report here the results of modification of carboxyls of fasciculin 2 from Dendrouspis angusticeps (eastern green mamba) by a method that neutralized the negative charge. (Preliminary results were presented at the 10th European Symposium on Animal, Plant and Microbial Toxins, Paris, 24 September 1992; Cerveiiansky et al., 1993). The toxin (for isolation see Dajas et al., 1992) was treated in 0.1 M cacodylate buffer (pH 5.2) for 1 hr at room temperature with norleucine methylester (90 times molar excess to carboxyls) and I-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (30 times molar excess) (Hoare and Koshland, 1967; Carraway and Koshland, 1972). After desalting on Sephadex G-25 in 1% acetic acid, the sample was freeze-dried and the derivatives were then isolated by HPLC on the cation-exchanger BioGel TSK SP-5-PW (sulphopropyl) (Fig. 2A). Prior to the ion-exchange a part of the sample was treated for 5 hr at room temprature with 0.5 M hydroxylamine (pH 7.0) to reverse the tyrosine modification (Carraway and Koshland, 1968) (Fig. 2B). Inhibitory activity was assayed with rat brain AChE in 50 mM sodium phosphate buffer, 0.025% Triton X-100 at 37°C as described elsewhere (Karlsson cl al., 1984).
*Author
to whom
correspondence
should
be addressed
718
Short
Communications
Fig. 1. Amino acid sequence of fasciculin 2 from Dendrouspis angusticeps (green mamba). Three fasciculins are known. Fasciculin 1, also from the green mamba, has Tyr instead of Asn at position 47, and fasciculin 3 from D. polylepis (black mamba) or D. uiridis (Western green mamba) has Be 2 instead of Met, Lys-Asp (15-16) instead of Thr-Asn and Tyr 47 instead of Asn (Marcbot et al., 1993). All fasciculins have a net charge of + 4 at neutral pH. Most of the cationic groups are in the first part of the molecule, residues l-37, and the peptide 2437 is very cationic with six positive and no negative charges. Five of the six or seven negative charges are in the last part, residues 45-61.
The elution of A-E was not affected by hydroxylamine (Fig. 2B), indicating modification only of carboxyl groups, while F-K were tyrosine derivatives. After treatment with hydroxylamine, F and G eluted as the native toxin and A, respectively, and they were derivatives of these compounds. The norleucine content (determined by amino acid analysis) of A-E varied between 0.94 (derivative D) and 1.15 (B) mole/mole toxin (mean 1.06 f 0.06), indicating that each derivative had one carboxyl modified. Five derivatives were obtained, and evidently one of the carboxyls of fasciculin 2 did not react. The I& (concentration to give 50% inhibition) of native toxin was 0.56 nM and of A-E between 0.45 (derivative E) and 0.65 nM (D) (mean 0.55 + 0.05 nM). The activities compared to that of the native toxin (I& of native toxin/ICso of derivative) were between 124 and 86%. Thus, modification had no significant effect and carboxyl groups do not participate in the binding to AChE. The derivatives A-E have lost one negative charge each and, as expected, they eluted from the cation-exchanger BioGel TSK SP after the native toxin. They have the same net charge, and the separation may depend on different distribution of charges created by the modification. Other fasciculin derivatives of the same net charge have also been separated by ion-exchange chromatography (Cerveiiansky et al., 1994,1995). Modification of tyrosine should not change the charge at neutral pH, but the tyrosine derivatives eluted after the carboxyl derivatives, even F which had no carboxyl modified. The matrix of BioGel TSK SP is a hydroxylated vinyl polymer and, besides electrostatic interactions, hydrophobic forces and perhaps also hydrogen bonding can influence the separation. Fasciculin 1 has Tyr 47 (hydrophobic) and fasciculin 2 Asn 47 (hydrophilic). Fasciculin 1 elutes after fasciculin 2, probably because of stronger hydrophobic interaction with the matrix, while they elute in reversed order from SP-Sephadex C-25, which has a more hydrophilic matrix. Norleucine methylester has a hydrophobic CHrCHrCHrCHrmoiety and coupling to a tyrosine residue probably results in strong hydrophobic interaction with the matrix and retarded elution. We have performed many experiments to identify the binding sites between fasciculin and AChE. Since this is the most recent of these experiments, it is perhaps appropriate to discuss them collectively here. Several observations indicate that fasciculins bind to a peripheral site of AChE: (1) propidium, a peripheral site ligand, is displaced from its binding site by fasciculin (Karlsson ef al., 1984); (2) affinity labelling of electric eel AChE at a peripheral site with N,N’-dimethyl-2-phenylaziridium in the presence of edrophonium to protect the active
120
Short Communications
B
,/
I
0
d FASP
FAS2
r (min)
70 0
(min)
70
Fig. 2. Isolation of fasciculin derivatives by HPLC on the cation-exchanger BioGel TSK SP-5-PW (7.5 x 75 mm) equilibrated with water. Buffer A: H20, B: 1.00 M ammonium acetate (pH 6.7). Programme: O-5 min elution with 15% B, 560 min, linear gradient to 60% B, 60-80 min linear gradient to 80% B. Flow: 0.7 ml/min. Fractions collected as indicated. (A) Before and (B) after treatment with hydroxylamine. Absorbance at full scale is 0.4 in A and 0.08 in B.
site made the enzyme about one million times less sensitive to fascicuhn. This was not due to great structural perturbations, since K, for acetylthiocholine increased only four times (from 0.085 to 0.37 mM) and the turnover number k,, did not change significantly, (3.5 f 0.5) for native and (2.8 k 0.2) x 10S/per min for modified AChE (Duran et al., 1994). The reagent introduces a positive charge, and probably labelled a tryptophan in the peripheral site (Weise et al., 1990); (3) site-directed mutagenesis of three aromatic amino acids (Y72, Yl24, W286) in the peripheral site of mouse AChE had a large effect. Fasciculin had a I00 times lower affinity for the mutant Y124Q, 4000 times lower for Y72N, and one million times lower for W286R (Radic et al., 1994). Since many peripheral site ligands, such as propidium and decamethonium, are cationic, the role of amino groups and arginine residues was studied by chemical modifications that neutralized the positive charge. The affinity of the monolabelled derivatives was reduced only by 60-85%, or 3- to 7-fold (Cervenansky et al., 1994, 1995). Cationic groups should bind to the peripheral site by interaction with the n-electrons of the aromatic amino acids (Burley and Petsko, 1986; Sussman and Silman, 1992). Modification of AChE, chemically or by mutagenesis, should have abolished the n-electron interaction, as would neutralization of the positive charges of amino groups and arginine residues of the toxin. A similar modification of a binding site in its toxin or enzyme part should have about the same effect; however, this was not the case, since the modification of the enzyme had a much greater effect. Therefore, it seems unlikely that the positive charge of any amino acid binds to AChE by interaction with the n-electrons of the aromatic amino acids in the peripheral site. Binding to this site may depend on hydrophobic interactions with the aromatic amino acids or hydrogen bonding between the n-electrons and hydrogen donor groups in fasciculin.
Acknowledgemenrs-This work was supported by the International Foundation for Science, Stockholm, the International Program in the Chemical Sciences, Uppsala University, and the Swedish Natural Science Research Council.
REFERENCES Burley A. L. and Petsko G. A. (1986) Amino-aromatic interactions in proteins. FEBS Lett. 203, 139143. Carraway K. L. and Koshland D. E. Jr (1968) Reaction of tyrosine residues in proteins with carbodiimide reagents. B&hem. biophys. Acta 160, 272-274.. Carraway K. L. and Koshland D. E. Jr (1972) Carbodiimide modification of proteins. Merh. Enzymol. 25, 616623.. Cervefiansky, C., Dajas, F., Harvey, A. L. and Karlsson, E. (1991) Fasciculins, anticholinesterase toxins from mamba venoms: biochemistry and pharmacology. In: International Encyclopedia of Pharmacology and Therapeutics: Snake Toxins, pp. 131-164 (Harvey, A. L., Ed.). New York: Pergamon Press.
Short
Communications
721
Cerveiiansky C., Dajas F., Engstrom A. and Karlsson E. (1993) Study of structure-activity relationship of fasciculin by chemical modification of arginine, lysine and carboxylate groups. Toxicon 31, 500. Cerveriansky C., Engstrom A. and Karlsson E. (1994) Study of structure-activity relationship of fasciculin by acetylation of amino groups. Biochem. biophys. Acta 1199, l-5. Cervefiansky C., Engstriim A. and Karlsson E. (1995) Role of arginine residues for the activity of fasciculin. Eur. J. Biochem. 229, 2X&275. Dajas F., Silveira R. and Cerveiiansky C. (1992) Fasciculin: neuropharmacology of a potent anticholinesterase polypeptide. Meth. Neurosci. 8, 258-270. DurPn R., Cervefiansky C., Dajas F. and Tipton K.F. (1994) Fasciculin inhibition of acetylcholinesterase is prevented by chemical modification of the enzyme at a peripheral site. Biochem. biophys. Acta 1201, 381-388. Hoare D. G. and Koshland D. E. Jr (1967) A method for quantitative modification and estimation of carboxylic acid groups in proteins. J. biol. Chem. 242, 2447-2453.. Karlsson E., Mbugua P. M. and Rodriguez-Ithurralde D. (1984) Fasciculins, anticholinesterase toxins from the venom of the green mamba Dendroaspis angusticeps. J. Physiol., Paris 79, 232-240. Marchot P., KhClif A., Ji Y. H., Mansuelle P. and Bougis P. (1993) Binding of ‘251-fasciculin to rat brain acetylcholinesterase. J. biol. Chem. 268, 12458-12467. Radic Z., Duran R., Vellom D. C., Li Y., Cerveiiansky C. and Taylor P. (1994) Site of fasciculin interaction with acetylcholinesterase. J. biol. Chem. 269, 11233-l 1239. Sussman J. L. and Silman I. (1992) Acetylcholinesterase: structure and use as a model for specific cation-protein interactions. Curr. Opin. Sfruet. Biol. 2, 721-729. Weise C., Kreienkamp H. J., Raba R., Pedak A., Aaviksaar A. and Hucho F. (1990) Anionic subsites of the acetylcholinesterase from Torpedo californica: affinity labelling with the cationic reagent N.N-dimethyl-2phenyl-aziridinium. EMBO J. 9, 3885-3888.
T&con. Vol. 34, No 6, pp. 718-721, 1996 Copyright 8 1996 Elsevier Science Ltd Pnnted in Great Britain. All rights reserved 0041~)101/96 $9.50 + 0.00
SHORT COMMUNICATION FASCICULIN: MODIFICATION OF CARBOXYL GROUPS AND DISCUSSION OF STRUCTURE-ACTIVITY RELATIONSHIP CARLOS CERVERANSKY,’ ROSARIO DURAN’ and EVERT KARLSSON’* ‘Institute de Investigaciones Biologicas Clemente Estable, Avenida ltalia 3318, 11600 Montevideo, Uruguay, and 2Department of Biochemistry, Biomedical Centre, Box 576, 751 23 Uppsala, Sweden. (Received
7 November
1995; Accepted
24 November
1995)
C. Cervenansky, R. Duran and E. Karlsson. Fasciculin: modification of carboxyl groups and discussion of structure-activity relationship. Toxicon 34, 7 18-72 1, 1996.-Norleucine methylester was coupled to carboxylates of fasciculin 2, a snake toxin that inhibits acetylcholinesterase (AChE). This neutralized negative charges but had no effect on the activity, suggesting that carboxyls do not participate in binding to AChE. Earlier results are discussed. Modification of three aromatic amino acids in the peripheral site of AChE, the binding site for fasciculin, decreased the affinity 100 to one million times. Neutralizing the charge of cationic groups of fasciculin lowered the affinity only three to seven times. A change in either the toxin or enzyme part of a binding site should have about the same effect. Since this was not so, it suggests that cationic groups of fasciculin do not bind to aromatic rings in the peripheral site. Copyright 0 1996 Elsevier Science Ltd
Fasciculins are cationic molecules with a very asymmetrical distribution of positive and negative charges (Cerveiiansky ef al., 1991) (Fig. 1). The role of cationic groups has been studied by chemical modification of amino groups and arginine residues (Cerveiiansky et al., 1994, 1995). We report here the results of modification of carboxyls of fasciculin 2 from Dendrouspis angusticeps (eastern green mamba) by a method that neutralized the negative charge. (Preliminary results were presented at the 10th European Symposium on Animal, Plant and Microbial Toxins, Paris, 24 September 1992; Cerveiiansky et al., 1993). The toxin (for isolation see Dajas et al., 1992) was treated in 0.1 M cacodylate buffer (pH 5.2) for 1 hr at room temperature with norleucine methylester (90 times molar excess to carboxyls) and I-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (30 times molar excess) (Hoare and Koshland, 1967; Carraway and Koshland, 1972). After desalting on Sephadex G-25 in 1% acetic acid, the sample was freeze-dried and the derivatives were then isolated by HPLC on the cation-exchanger BioGel TSK SP-5-PW (sulphopropyl) (Fig. 2A). Prior to the ion-exchange a part of the sample was treated for 5 hr at room temprature with 0.5 M hydroxylamine (pH 7.0) to reverse the tyrosine modification (Carraway and Koshland, 1968) (Fig. 2B). Inhibitory activity was assayed with rat brain AChE in 50 mM sodium phosphate buffer, 0.025% Triton X-100 at 37°C as described elsewhere (Karlsson cl al., 1984).
*Author
to whom
correspondence
should
be addressed
718
Short
Communications
Fig. 1. Amino acid sequence of fasciculin 2 from Dendrouspis angusticeps (green mamba). Three fasciculins are known. Fasciculin 1, also from the green mamba, has Tyr instead of Asn at position 47, and fasciculin 3 from D. polylepis (black mamba) or D. uiridis (Western green mamba) has Be 2 instead of Met, Lys-Asp (15-16) instead of Thr-Asn and Tyr 47 instead of Asn (Marcbot et al., 1993). All fasciculins have a net charge of + 4 at neutral pH. Most of the cationic groups are in the first part of the molecule, residues l-37, and the peptide 2437 is very cationic with six positive and no negative charges. Five of the six or seven negative charges are in the last part, residues 45-61.
The elution of A-E was not affected by hydroxylamine (Fig. 2B), indicating modification only of carboxyl groups, while F-K were tyrosine derivatives. After treatment with hydroxylamine, F and G eluted as the native toxin and A, respectively, and they were derivatives of these compounds. The norleucine content (determined by amino acid analysis) of A-E varied between 0.94 (derivative D) and 1.15 (B) mole/mole toxin (mean 1.06 f 0.06), indicating that each derivative had one carboxyl modified. Five derivatives were obtained, and evidently one of the carboxyls of fasciculin 2 did not react. The I& (concentration to give 50% inhibition) of native toxin was 0.56 nM and of A-E between 0.45 (derivative E) and 0.65 nM (D) (mean 0.55 + 0.05 nM). The activities compared to that of the native toxin (I& of native toxin/ICso of derivative) were between 124 and 86%. Thus, modification had no significant effect and carboxyl groups do not participate in the binding to AChE. The derivatives A-E have lost one negative charge each and, as expected, they eluted from the cation-exchanger BioGel TSK SP after the native toxin. They have the same net charge, and the separation may depend on different distribution of charges created by the modification. Other fasciculin derivatives of the same net charge have also been separated by ion-exchange chromatography (Cerveiiansky et al., 1994,1995). Modification of tyrosine should not change the charge at neutral pH, but the tyrosine derivatives eluted after the carboxyl derivatives, even F which had no carboxyl modified. The matrix of BioGel TSK SP is a hydroxylated vinyl polymer and, besides electrostatic interactions, hydrophobic forces and perhaps also hydrogen bonding can influence the separation. Fasciculin 1 has Tyr 47 (hydrophobic) and fasciculin 2 Asn 47 (hydrophilic). Fasciculin 1 elutes after fasciculin 2, probably because of stronger hydrophobic interaction with the matrix, while they elute in reversed order from SP-Sephadex C-25, which has a more hydrophilic matrix. Norleucine methylester has a hydrophobic CHrCHrCHrCHrmoiety and coupling to a tyrosine residue probably results in strong hydrophobic interaction with the matrix and retarded elution. We have performed many experiments to identify the binding sites between fasciculin and AChE. Since this is the most recent of these experiments, it is perhaps appropriate to discuss them collectively here. Several observations indicate that fasciculins bind to a peripheral site of AChE: (1) propidium, a peripheral site ligand, is displaced from its binding site by fasciculin (Karlsson ef al., 1984); (2) affinity labelling of electric eel AChE at a peripheral site with N,N’-dimethyl-2-phenylaziridium in the presence of edrophonium to protect the active
120
Short Communications
B
,/
I
0
d FASP
FAS2
r (min)
70 0
(min)
70
Fig. 2. Isolation of fasciculin derivatives by HPLC on the cation-exchanger BioGel TSK SP-5-PW (7.5 x 75 mm) equilibrated with water. Buffer A: H20, B: 1.00 M ammonium acetate (pH 6.7). Programme: O-5 min elution with 15% B, 560 min, linear gradient to 60% B, 60-80 min linear gradient to 80% B. Flow: 0.7 ml/min. Fractions collected as indicated. (A) Before and (B) after treatment with hydroxylamine. Absorbance at full scale is 0.4 in A and 0.08 in B.
site made the enzyme about one million times less sensitive to fascicuhn. This was not due to great structural perturbations, since K, for acetylthiocholine increased only four times (from 0.085 to 0.37 mM) and the turnover number k,, did not change significantly, (3.5 f 0.5) for native and (2.8 k 0.2) x 10S/per min for modified AChE (Duran et al., 1994). The reagent introduces a positive charge, and probably labelled a tryptophan in the peripheral site (Weise et al., 1990); (3) site-directed mutagenesis of three aromatic amino acids (Y72, Yl24, W286) in the peripheral site of mouse AChE had a large effect. Fasciculin had a I00 times lower affinity for the mutant Y124Q, 4000 times lower for Y72N, and one million times lower for W286R (Radic et al., 1994). Since many peripheral site ligands, such as propidium and decamethonium, are cationic, the role of amino groups and arginine residues was studied by chemical modifications that neutralized the positive charge. The affinity of the monolabelled derivatives was reduced only by 60-85%, or 3- to 7-fold (Cervenansky et al., 1994, 1995). Cationic groups should bind to the peripheral site by interaction with the n-electrons of the aromatic amino acids (Burley and Petsko, 1986; Sussman and Silman, 1992). Modification of AChE, chemically or by mutagenesis, should have abolished the n-electron interaction, as would neutralization of the positive charges of amino groups and arginine residues of the toxin. A similar modification of a binding site in its toxin or enzyme part should have about the same effect; however, this was not the case, since the modification of the enzyme had a much greater effect. Therefore, it seems unlikely that the positive charge of any amino acid binds to AChE by interaction with the n-electrons of the aromatic amino acids in the peripheral site. Binding to this site may depend on hydrophobic interactions with the aromatic amino acids or hydrogen bonding between the n-electrons and hydrogen donor groups in fasciculin.
Acknowledgemenrs-This work was supported by the International Foundation for Science, Stockholm, the International Program in the Chemical Sciences, Uppsala University, and the Swedish Natural Science Research Council.
REFERENCES Burley A. L. and Petsko G. A. (1986) Amino-aromatic interactions in proteins. FEBS Lett. 203, 139143. Carraway K. L. and Koshland D. E. Jr (1968) Reaction of tyrosine residues in proteins with carbodiimide reagents. B&hem. biophys. Acta 160, 272-274.. Carraway K. L. and Koshland D. E. Jr (1972) Carbodiimide modification of proteins. Merh. Enzymol. 25, 616623.. Cervefiansky, C., Dajas, F., Harvey, A. L. and Karlsson, E. (1991) Fasciculins, anticholinesterase toxins from mamba venoms: biochemistry and pharmacology. In: International Encyclopedia of Pharmacology and Therapeutics: Snake Toxins, pp. 131-164 (Harvey, A. L., Ed.). New York: Pergamon Press.
Short
Communications
721
Cerveiiansky C., Dajas F., Engstrom A. and Karlsson E. (1993) Study of structure-activity relationship of fasciculin by chemical modification of arginine, lysine and carboxylate groups. Toxicon 31, 500. Cerveriansky C., Engstrom A. and Karlsson E. (1994) Study of structure-activity relationship of fasciculin by acetylation of amino groups. Biochem. biophys. Acta 1199, l-5. Cervefiansky C., Engstriim A. and Karlsson E. (1995) Role of arginine residues for the activity of fasciculin. Eur. J. Biochem. 229, 2X&275. Dajas F., Silveira R. and Cerveiiansky C. (1992) Fasciculin: neuropharmacology of a potent anticholinesterase polypeptide. Meth. Neurosci. 8, 258-270. DurPn R., Cervefiansky C., Dajas F. and Tipton K.F. (1994) Fasciculin inhibition of acetylcholinesterase is prevented by chemical modification of the enzyme at a peripheral site. Biochem. biophys. Acta 1201, 381-388. Hoare D. G. and Koshland D. E. Jr (1967) A method for quantitative modification and estimation of carboxylic acid groups in proteins. J. biol. Chem. 242, 2447-2453.. Karlsson E., Mbugua P. M. and Rodriguez-Ithurralde D. (1984) Fasciculins, anticholinesterase toxins from the venom of the green mamba Dendroaspis angusticeps. J. Physiol., Paris 79, 232-240. Marchot P., KhClif A., Ji Y. H., Mansuelle P. and Bougis P. (1993) Binding of ‘251-fasciculin to rat brain acetylcholinesterase. J. biol. Chem. 268, 12458-12467. Radic Z., Duran R., Vellom D. C., Li Y., Cerveiiansky C. and Taylor P. (1994) Site of fasciculin interaction with acetylcholinesterase. J. biol. Chem. 269, 11233-l 1239. Sussman J. L. and Silman I. (1992) Acetylcholinesterase: structure and use as a model for specific cation-protein interactions. Curr. Opin. Sfruet. Biol. 2, 721-729. Weise C., Kreienkamp H. J., Raba R., Pedak A., Aaviksaar A. and Hucho F. (1990) Anionic subsites of the acetylcholinesterase from Torpedo californica: affinity labelling with the cationic reagent N.N-dimethyl-2phenyl-aziridinium. EMBO J. 9, 3885-3888.