- Email: [email protected]
Study of structure-activity relationship of fasciculin by acetylation of amino groups
1
Biochimica et Biophysica Acta, 1199 (1994) 1-5 © 1994 Elsevier Science B.V. All rights reserved 0304-4165/94/$07.00
BBAGEN 23859
Study of structure-activity relationship of fasciculin by acetylation of amino groups v
~
p
O
Carlos Cervenanslc9 a, Ake Engstr6m b and Evert Karlsson e,, Instituto de Investigaciones Biol6gicas Clemente Estable, Av. Italia 3318, 11600 Montevideo (Uruguay), b Department of Immunology, Biomedical Centre, Box 582, 751 23 Uppsala (Sweden) and ¢ Department of Biochemistry, Biomedical Centre, Box 576, 751 23 Uppsala (Sweden) (Received 10 March 1993) (Revised manuscript received 14 June 1993)
Key words: Anticholinesterase toxin; Fasciculin; Modification; Amino groups; (Green mamba)
Dendroaspis angusticeps (green mamba) has two toxins, fasciculins, that are non-competitive inhibitors of acetylcholinesterase. Amino groups of fasciculin 2 were acetylated with acetic anhydride. The monoacetyl derivatives of the e-amino groups (Lys 25, 32, 51 and 58) retained between 28 and 43% of the initial activity and that of the a-amino group 72%. Acetylation of Lys 25 that has the most reactive amino group decreased the activity by 65% apparently without producing structural perturbations, since the circular dichroism spectrum was not affected. The three-dimensional structure shows a cationic cluster formed by Lys 32, 51, Arg 24 and 28. A comparison of 175 sequences of homologous toxins shows that Lys 32 is unique for fasciculin. Acetylation of lysine residues in the cluster had a large effect and reduced the activity by 72% (Lys 32) and 57% (Lys 51). This suggests an important role for the cationic cluster. Lys 25 together with Lys 32 and 51 were, therefore, assumed to be in the active site.
Introduction The West African snake Dendroaspis angusticeps (green m a m b a ) has two toxins called fasciculins because they produce long lasting fasciculations. They are non-competitive inhibitors of acetylcholinesterase (ACHE) [1,2]. Cholinesterases have a very different sensitivity towards fasciculins. AChEs from rat brain, human erythrocytes and electroplax of electric eel Electrophorus electricus are inhibited with Ki about 10 -11 M [3-5] and pseudocholinesterases as human serum cholinesterase with Ki about 0.5/zM. A second group of enzymes is partially (10-30%) inhibited by low ( < 0.5 nM) fasciculin, increasing the toxin concentration to about 1 nM re-activates the enzymes to 90-110% of their initial activity. AChEs from guinea pig ileum, ventricle and uterus behave in that way. A third group consists of enzymes insensitive to fasciculin; AChEs from chick biventer cervicis muscle and brain and from insects, heads of Musca domestica (common house fly) and cobra (Naja naja) venom [3,4]. Fasciculins displace propidium from its binding site on ACHE. Since propidium is a probe for a peripheral anionic site, it is concluded that fasciculins also bind to the same site [5]. The different sensitivity of cholin-
* Corresponding author. Fax: +46 18 552139.
SSDI 0 3 0 4 - 4 1 6 5 ( 9 3 ) E 0 0 6 7 - S
esterases to fasciculin should depend on the nature of their peripheral sites. Fasciculins are basic proteins of 61 amino acid residues and four disulphides. A large number of A C h E inhibitors are cations, e.g. neostigmine, physostigmine and propidium. Therefore, it is logical to investigate the role of cationic groups for the activity of fasciculin. In this article we describe the modification of amino groups by acetylation with acetic anhydride and its effect on the activity. Materials and methods
Isolation of fasciculins The fasciculins were isolated from the green m a m b a venom by gel filtration on Sephadex G-50 and ion-exchange chromatography on Bio-Rex 70 [5]. The final purification was obtained by H P L C ion-exchange on a TSK-SP-5PW cation-exchanger (SP = sulphopropyl) using a linear gradient of ammonium acetate (pH 6.8). The fasciculins account for 3 - 6 % of the venom protein and the relative proportion between fasciculin 1 and 2 is about 1/3. Fasciculin 2 was used in the experiments.
Acetylation and isolation of derivatives Amino groups were acetylated with acetic anhydride. A five-fold molar excess of amino groups over anhydride was used to obtain predominantly mono-
acetyl derivatives. In one experiment radioactive (tritiated) acetic anhydride (Amersham) was used. The reaction was carried out at room temperature in 0.05 M N-ethyl-morpholine acetate (pH 7.5) and was stopped after 40 min by gel filtration on Sephadex G-25 in 1% acetic acid. The protein fraction was freeze-dried and divided into two parts. The first one was fractionated by HPLC on the cation-exchanger TSK-SP-5PW (7.5 × 75 mm) using a linear gradient of ammonium acetate (pH 6.8). The peaks were pooled, freeze-dried and rerun on TSK-SP-5PW to remove background contamination. Each peak was also submitted to a HPLC reversed phase chromatography on a C4 column using a gradient of 0.1% trifluoroacetic acid (pH 2.3) vs. 0.08% trifluoroacetic acid in 60% acetonitrile. The second part was treated with 1 M hydroxylamine (pH 7.0) [5] for 6 h to deacetylate tyrosine [5], desalted, freeze-dried and submitted to ion-exchange chromatography on TSK-SP-5PW as above. The degree of acetylation was obtained as the ratio r a d i o a c t i v i t y / p r o t e i n concentration expressed as cpm/A276 (absorbance at 276 nm) and confirmed by mass spectrometry.
Inhibitory activity of acetylated derivatives Acetylcholinesterase from the electroplax of Electrophorus electricus (electric eel) (Sigma, Product number C 2888) was used and the enzyme activity was assayed spectrophotometrically with 1 mM acetylthiocholine as substrate and 0.2 mM 4,4'-dithiopyridine as chromophore [6]. The IC50 (concentration to produce 50% inhibition) was determined for each derivative.
Other pyridylethyl derivatives were digested with lysyl endopeptidase (Wako Chemicals, Dallas, TX, USA): 50 mM Tris-HC1, 3 M guanidinium hydrochloride (to solubilize sample) (pH 8.5), molar ratio enz y m e / substrate 1/200, 37°C and 6 h. The digests were submitted to reversed phase chromatography on a C18 column in the same system as earlier. The peptides were then analysed by plasma desorption mass spectrometry. Results
Isolation of acetylated derivatives The isolation of the acetyl derivatives by HPLC ion-exchange chromatography is shown in Fig. 1. The peaks A to J were pooled, freeze-dried, dissolved in water and the absorbance at 276 nm was measured. Aliquots were taken for determination of the radioactivity. The ratio cpm/A276 for peaks A, B, and D was 80000 + 8000 cpm and for peaks C, E, F, H and I 4 0 0 0 0 + 5000 cpm (Fig. 1) representing di- and monoacetyl derivatives, respectively. The last peak J eluted in the same position as the native toxin. It contained low radioactivity (cpm/A276 = 1900) due to trailing from the preceding peaks and rechromatography removed practically all of it. The peaks A to I were rechromatographed and their activity was determined and expressed as IC50 for the native toxin/ICs0 for the derivative. IC50 for the native toxin was 0.13 nM (Table I). Ion-exchange chromatography of a sample treated with 1 M hydroxylamine (pH 7.0) gave a chromatogramme with the same number of peaks as in Fig. 1, nor was any peak signifi-
Localization of acetyl groups Acetyl derivatives were reduced with dithiothreitol and alkylated with 4-vinylpyridine as described by Hermodson et al. [7]. The resulting pyridylethyl derivatives were desalted by HPLC on a C4 reversed phase column using a gradient of 0.1% trifluoroacetic acid versus 0.08% trifluoroacetic acid in 60% acetonitrile and freeze-dried. The pyridylethyl derivative containing the most reactive amino group (the largest peak in ion-exchange chromatography of the acetyl derivatives) was digested with V8-protease: 0.1 M ammonium bicarbonate (pH 7.9), molar ratio e n z y m e / s u b s t r a t e 1/25, 37°C and 2 h. The peptide mixture was analyzed by plasma desorption mass spectrometry [8] with a Bioion 20 mass spectrometer (Applied Biosystems AB, Uppsala, Sweden). The peptides were isolated on a C18 reversed phase column in the same system as above. The peptide containing the acetyl group was submitted to sequence determination by Edman degradation with an automatic gas phase sequencer (Applied Biosystems Protein Sequencer 470A) and on line analysis of the phenylthiohydantoin derivatives with an Applied Biosystems 120A P T H Analyzer.
0.50-
G25r,.)
IZ
F
0
~00
i
i
!
|
|
0
20
40
60
80
RETENTION
TIME
i
100 (rain)
Fig. 1. Isolation of acetylated derivatives of fasciculin 2 by H P L C ion-exchange chromatography on a TSK-SP-5PW column ( 7 . 5 x 7 5 mm) equilibrated with water. Freeze-dried sample was dissolved in H 2 0 and applied to the column. After sample application the column was eluted for 4 rain with H 2 0 and then for 80 min with a linear gradient of H 2 0 vs. 1 M a m m o n i u m acetate (pH 6.8). The concentration of acetate increased by 2.5 m M (0.25%) per min. Flow rate 0.7 m l / m i n .
TABLE I
Acetyl derivatives of fasciculin 2 Derivative (numbering as in Fig. 1)
Acetyl groups per molecule
Remaining activity a %
Acetyl in position
A B C D E F G H I J
2b 2 2 2 1 1 1c 1 1 0
8 15 11 21 28 40 43 35 72 100
not determined Lys 58, Lys 25? not determined Lys 25, N-terminus Lys 32 Lys 58 Lys 51 Lys 25 N-terminus native toxin
a IC50 for native toxin = 0.13 nM. Remaining activity = IC50 for native t o x i n / I C s o for derivative. b Determined only from ratio cpm/A276. c Determined only by mass spectrometry.
cantly smaller. This indicates that tyrosine residues were not acetylated.
Localization of the acetyl groups 1. Derivative Ho The pyridylethyl derivative was digested with V8 protease that cleaves with high specificity at the C-terminal side of Glu residues. Cleavage should occur after GIu 19 and Glu 48 (Fig. 2). Three fragments with the masses 1675, 2297 and 3702 Da were detected by mass spectrometry. The two smallest fragments were the peptides 1-19 and 50-61, since the masses obtained by mass spectrometry agree very well with the composition data, 1676 and 2299 Da, respectively. The largest fragment had a mass 43 Da higher than for the peptide 20-48 and contained evidently an
+
40
-
_
2. The primary structure of fasciculin 2 (20). The toxin has an unusual distribution of charges. T h e second loop is very cationic and the third loop anionic. Fig.
acetyl group (42 Da) either on the side chain of Lys 25 or Lys 32. The peptide 20-48 was isolated by reversed phase chromatography on a C4 column and submitted to Edman degradation. The following sequence was obtained: Asn-Ser-Cys-Tyr-Arg-?-Ser-Arg-Arg-His-ProPro-Lys-Met-Val-Leu-Gly-Arg. To confirm that position 6 contained an acetylated Lys, a peptide Ala-AlaLys-Leu was synthesized by the Merrifield method and the side chain of Lys was reacted with acetic anhydride before the cleavage of the peptide from the resin. The phenylthiohydantoin derivative at step 3 in the Edman degradation eluted with the same retention time as the phenylthiohydantoin derivative at position 6 of the fragment 20-48. Acetylation of Lys 25 decreased the activity by 65% (Table 1). Circular dichroism spectra of derivative H and native fasciculin 2 were identical indicating that acetylation did not have a large effect on the conformation. 2. Derivatives E, F, G and I. To confirm the specificity of lysylendopeptidase, the pyridylethyl-derivative of H was digested with this enzyme and peptides were isolated by reversed phase chromatography. The peptides 1-32, 33-51 and 52-58 were detected by mass spectrometry but not the peptide 25-32 or the C-terminal tripeptide. The enzyme split as expected after Lys residues but not if the side chain was acetylated. The sample for mass spectrometry is applied on a filter that is washed to remove excess salt and during that process the C-terminal tripeptide was probably lost. The pyridylethyl derivatives of E, F, G and I were digested with lysyl endopeptidase, the peptides were isolated and their molecular weights were determined by mass spectrometry. The relative amounts of monoacetyl derivatives reflect the reactivity of the amino groups. Acetyl-Lys 25 accounted for 75%, acetyl-Lys 58 for 14% and the derivatives of the N-terminus, Lys 32 and 51 each for 3-4%. 3. Derivatives B and D. The diacetyl derivatives B and E were treated in the same way. Derivative B had the acetyl groups on Lys 58 and probably on Lys 25 and D on the N-terminus and Lys 25 (Table I). Discussion The binding between fasciculin and acetylcholinesterase is strong as indicated by a Ki of about 10 -H M [4-6]. This should result from interaction of several amino acid residues in the toxin with the enzyme. A modification of one of these residues should not abolish but significantly decrease the activity [9,10]. But the decrease in activity can also depend on structural perturbations caused by the modification.
Acetylation of the four e-amino groups reduced the activity between 57% (Lys 51) and 72% (Lys 32) while acetylation of the a-amino group had less effect and decreased the activity only by 28%. But as stated above the drop in activity can have two reasons. It is likely, however, that the lower activity of some of the monacetyl derivatives depends on modification of an amino group in the active site and for others on structural perturbations. Acetylation of Lys 25 reduced the activity by 65% without producing any structural changes, since the circular dichroism spectrum was not affected. Therefore, Lys 25 is probably in the active site. Indirect evidence for the role of Lys 32 and 51 can be obtained from a comparison of sequences of homologous toxins and from the three-dimensional structure of fasciculin [10]. A large number of snake toxins are homologous to fasciculin, such as a-neurotoxins which bind to the nicotinic acetylcholine receptor and cardiotoxins that increase the permeability of biological membranes. At the present, sequences of 175 homologues are known [11], A homology alignment with Cys-residues in the same positions shows that Lys 32 is unique for fasciculins and it should therefore be of special importance. Lysine in positions corresponding to 25, 51 and 58 are found frequently in many toxins. However, an amino acid that is located in the same position in different types of toxins can still be part of the active site. This has been shown for oxiana II, a short a-neurotoxin from Naja naja oxiana. By using spin labeled, fluorescent and photoaffinity derivatives it was possible to show that this lysine residue interacts with the acetylcholine receptor [12,13]. The three dimensional structure of fasciculin 1 has been determined recently [10]. The two fasciculins differ only at position 47 where fasciculin 1 has tyrosine and fasciculin 2 asparagine. Fasciculin 1 has a similar overall structure as short a-neurotoxins [14,15] and cardiotoxins [16], a dense core with disulphide bridges and the three loops Cys 3 to Cys 22, Cys 22 to Cys 39 and Cys 52 long and extended as the central fingers of a hand. Lys 25 is on the opposite side of the toxin molecule compared Lys 32 and 51 which form a cationic quartet together with Arg 24 and 28. The side chain of Lys 58 points in 90 ° angle from that of Lys 32 (Fig. 3). The cluster of cationic groups probably has a functional role as already suggested [10]. It contains Lys 32 which is unique for fasciculin. Acetylation of both lysine residues had a large effect decreasing the activity by 72% (Lys 32) and 57% (Lys 51). Thus, indirect evidence from chemical modification and structural data suggest that Lys 32 and 51 have a functional role. The evidence for Lys 25 is based on chemical modification and circular dichroism spec-
K25
K32 K51 Fig. 3. Schematic ribbon drawing of the three-dimensional structure of fasciculin 1, with E-structures indicated by arrows. The molecule has no a-helices. The side chain of Lys 58 points into the plane of the figure and the side chains of the other lysine residues are approximately in the plane of the figure. The data programme MOLSCRIPT (21) was used for drawing the figure.
troscopy. Functionally important amino acids should be then located on both sides of the fasciculin molecule (Fig. 3). Acetylation of the a-amino group and Lys 58 was also followed by loss of activity, but structural data cannot give any further indication of their possible functional role. Lys 25 has the most reactive amino group, acetyl-Lys 25 accounts for 75% of the monoacetyl derivatives. The reactivity of an amino group depends on its pKa and accessibility [17] and these factors seem to combine in fasciculin to give Lys 25 its superreactivity. The corresponding Lys in other toxins can also be very reactive as in the short a-neurotoxin erabutoxin b where it has the most reactive amino group [18]. The three-dimensional structure of AChE from Torpedo californica has been determined [19]. The active site is at the bottom of a deep gorge and the peripheral anionic site with the binding site for propidium [16] at the rim of the gorge. Since fasciculin and propidium have a common binding site [5], the toxin should also bind to the rim.
Acknowledgements This work was supported by the Swedish National Science Research Council, by O.E. och Edla Johanssons Stiftelse, by the International Program in the Chemical Sciences (IPICS), Uppsala University, and by the International Foundation for Science (IFS), Stockholm. We thank Dr. Andr~ M6nez for recording of the circular dichroism spectra, Dr. Juan Fontecilla-Camps for information on the three-dimensional structure of
fasciculin and Dr. Bengt Westerlund for computer drawing of the structure. References 1 Rodrfguez-Ithurralde, D., Silveira, R., Barbeito, L. and Dajas, F. (1983) Neurochem, Int. 5, 267-274. 2 Karlsson, E., Mbugua, P.M. and Rodr~guez-Ithurralde, D. (1985) Pharmac. Ther. 30, 259-276. 3 Cervefiansky, C., Dajas, F., Harvey, A.L. and Karlsson, E. (1991) In: International Encyclopedia of Pharmacology and Therapeutics: Snake Toxins (Harvey, A.L., ed.), pp. 131-164, Pergamon Press, New York. 4 Puu, G. and Koch, M. (1990) Biochem. Pharmacol. 40, 2209-2214. 5 Riordan, J.F. and Vallee, B.L. (1967) Methods Enzymol. 11, 570-576. 6 Karlsson, E., Mbugua, P.M. and Rodr~guez-lthurralde, D. (1984) J. Physiol. Paris 79, 232-240. 7 Hermodson, M.A., Ericsson, L.H., Neurath, H. and Walsh, K.A. (1973) Biochemistry 12, 3146-3153. 8 Sundqvist, B. and MacFarlane, R.D. (1985) Mass Spectrom. Rev. 4, 421-460. 9 Karlsson, E. and Sundelin, J. (1976) Toxicon 14, 295-306.
10 Le Du, M.H., Marchot, P., Bougis, P. and FonteciUa-Camps, J. (1992) J. Biol. Chem. 267, 22122-22130. 11 Mebs, D. and Claus, I. (1991) In: International Encyclopedia of Pharmacology and Therapeutics. Snake Toxins (Harvey, A.L., ed.), pp. 425-447. 12 Tsetlin, V.I., Karlsson, E., Arseniev, A.S., Utkin, Yu.N., Surin, A.M., Pashkov, V.S., Pluzhnikov, K.A., Ivanov, V.T., Bystrov, V.F. and Ovchinnikov, Yu. A. (1979) FEBS Lett. 106, 47-52. 13 Kreienkamp, H.J., Utkin, Yu.N., Weise, C., Machold, J., Tsetlin, V.I. and Hucho, F. (1992) Biochemistry 31, 8239-8244. 14 Tsernoglu, D. and Petsko, G.A. (1976) FEBS Lett. 68, 1-4. 15 Low, B., Preston, H.S., Sato, A., Rosen, L.S., Searl, J.E., Rudko, A.D. and Richardson, J.S. (1976) Proc. Natl. Acad. Sci. USA. 73, 2991-2994. 16 Smith, J.L., Corfield, P.W.R., Hendrickson, W.A. and Low, B. (1988) Acta Cryst. A44, 357-368. 17 Bosshard, H.R. (1979) Methods Biochem. Anal. 25, 273-301. 18 Hori, H. and Tamiya, N. (1976) Biochem. J. 153, 217-222. 19 Sussman, J.l., Harel, M., Frolow, F., Oefner, C., Goldman, A., Toker, L. and Silman, I. (1991) Science 253, 872-879. 20 Viljoen, C.C. and Botes, D.P. (1973) J. Biol. Chem. 248, 49154919. 21 Kraulis, P. (1991) J. Appl. Crystallogr. 24, 946-950.
Biochimica et Biophysica Acta, 1199 (1994) 1-5 © 1994 Elsevier Science B.V. All rights reserved 0304-4165/94/$07.00
BBAGEN 23859
Study of structure-activity relationship of fasciculin by acetylation of amino groups v
~
p
O
Carlos Cervenanslc9 a, Ake Engstr6m b and Evert Karlsson e,, Instituto de Investigaciones Biol6gicas Clemente Estable, Av. Italia 3318, 11600 Montevideo (Uruguay), b Department of Immunology, Biomedical Centre, Box 582, 751 23 Uppsala (Sweden) and ¢ Department of Biochemistry, Biomedical Centre, Box 576, 751 23 Uppsala (Sweden) (Received 10 March 1993) (Revised manuscript received 14 June 1993)
Key words: Anticholinesterase toxin; Fasciculin; Modification; Amino groups; (Green mamba)
Dendroaspis angusticeps (green mamba) has two toxins, fasciculins, that are non-competitive inhibitors of acetylcholinesterase. Amino groups of fasciculin 2 were acetylated with acetic anhydride. The monoacetyl derivatives of the e-amino groups (Lys 25, 32, 51 and 58) retained between 28 and 43% of the initial activity and that of the a-amino group 72%. Acetylation of Lys 25 that has the most reactive amino group decreased the activity by 65% apparently without producing structural perturbations, since the circular dichroism spectrum was not affected. The three-dimensional structure shows a cationic cluster formed by Lys 32, 51, Arg 24 and 28. A comparison of 175 sequences of homologous toxins shows that Lys 32 is unique for fasciculin. Acetylation of lysine residues in the cluster had a large effect and reduced the activity by 72% (Lys 32) and 57% (Lys 51). This suggests an important role for the cationic cluster. Lys 25 together with Lys 32 and 51 were, therefore, assumed to be in the active site.
Introduction The West African snake Dendroaspis angusticeps (green m a m b a ) has two toxins called fasciculins because they produce long lasting fasciculations. They are non-competitive inhibitors of acetylcholinesterase (ACHE) [1,2]. Cholinesterases have a very different sensitivity towards fasciculins. AChEs from rat brain, human erythrocytes and electroplax of electric eel Electrophorus electricus are inhibited with Ki about 10 -11 M [3-5] and pseudocholinesterases as human serum cholinesterase with Ki about 0.5/zM. A second group of enzymes is partially (10-30%) inhibited by low ( < 0.5 nM) fasciculin, increasing the toxin concentration to about 1 nM re-activates the enzymes to 90-110% of their initial activity. AChEs from guinea pig ileum, ventricle and uterus behave in that way. A third group consists of enzymes insensitive to fasciculin; AChEs from chick biventer cervicis muscle and brain and from insects, heads of Musca domestica (common house fly) and cobra (Naja naja) venom [3,4]. Fasciculins displace propidium from its binding site on ACHE. Since propidium is a probe for a peripheral anionic site, it is concluded that fasciculins also bind to the same site [5]. The different sensitivity of cholin-
* Corresponding author. Fax: +46 18 552139.
SSDI 0 3 0 4 - 4 1 6 5 ( 9 3 ) E 0 0 6 7 - S
esterases to fasciculin should depend on the nature of their peripheral sites. Fasciculins are basic proteins of 61 amino acid residues and four disulphides. A large number of A C h E inhibitors are cations, e.g. neostigmine, physostigmine and propidium. Therefore, it is logical to investigate the role of cationic groups for the activity of fasciculin. In this article we describe the modification of amino groups by acetylation with acetic anhydride and its effect on the activity. Materials and methods
Isolation of fasciculins The fasciculins were isolated from the green m a m b a venom by gel filtration on Sephadex G-50 and ion-exchange chromatography on Bio-Rex 70 [5]. The final purification was obtained by H P L C ion-exchange on a TSK-SP-5PW cation-exchanger (SP = sulphopropyl) using a linear gradient of ammonium acetate (pH 6.8). The fasciculins account for 3 - 6 % of the venom protein and the relative proportion between fasciculin 1 and 2 is about 1/3. Fasciculin 2 was used in the experiments.
Acetylation and isolation of derivatives Amino groups were acetylated with acetic anhydride. A five-fold molar excess of amino groups over anhydride was used to obtain predominantly mono-
acetyl derivatives. In one experiment radioactive (tritiated) acetic anhydride (Amersham) was used. The reaction was carried out at room temperature in 0.05 M N-ethyl-morpholine acetate (pH 7.5) and was stopped after 40 min by gel filtration on Sephadex G-25 in 1% acetic acid. The protein fraction was freeze-dried and divided into two parts. The first one was fractionated by HPLC on the cation-exchanger TSK-SP-5PW (7.5 × 75 mm) using a linear gradient of ammonium acetate (pH 6.8). The peaks were pooled, freeze-dried and rerun on TSK-SP-5PW to remove background contamination. Each peak was also submitted to a HPLC reversed phase chromatography on a C4 column using a gradient of 0.1% trifluoroacetic acid (pH 2.3) vs. 0.08% trifluoroacetic acid in 60% acetonitrile. The second part was treated with 1 M hydroxylamine (pH 7.0) [5] for 6 h to deacetylate tyrosine [5], desalted, freeze-dried and submitted to ion-exchange chromatography on TSK-SP-5PW as above. The degree of acetylation was obtained as the ratio r a d i o a c t i v i t y / p r o t e i n concentration expressed as cpm/A276 (absorbance at 276 nm) and confirmed by mass spectrometry.
Inhibitory activity of acetylated derivatives Acetylcholinesterase from the electroplax of Electrophorus electricus (electric eel) (Sigma, Product number C 2888) was used and the enzyme activity was assayed spectrophotometrically with 1 mM acetylthiocholine as substrate and 0.2 mM 4,4'-dithiopyridine as chromophore [6]. The IC50 (concentration to produce 50% inhibition) was determined for each derivative.
Other pyridylethyl derivatives were digested with lysyl endopeptidase (Wako Chemicals, Dallas, TX, USA): 50 mM Tris-HC1, 3 M guanidinium hydrochloride (to solubilize sample) (pH 8.5), molar ratio enz y m e / substrate 1/200, 37°C and 6 h. The digests were submitted to reversed phase chromatography on a C18 column in the same system as earlier. The peptides were then analysed by plasma desorption mass spectrometry. Results
Isolation of acetylated derivatives The isolation of the acetyl derivatives by HPLC ion-exchange chromatography is shown in Fig. 1. The peaks A to J were pooled, freeze-dried, dissolved in water and the absorbance at 276 nm was measured. Aliquots were taken for determination of the radioactivity. The ratio cpm/A276 for peaks A, B, and D was 80000 + 8000 cpm and for peaks C, E, F, H and I 4 0 0 0 0 + 5000 cpm (Fig. 1) representing di- and monoacetyl derivatives, respectively. The last peak J eluted in the same position as the native toxin. It contained low radioactivity (cpm/A276 = 1900) due to trailing from the preceding peaks and rechromatography removed practically all of it. The peaks A to I were rechromatographed and their activity was determined and expressed as IC50 for the native toxin/ICs0 for the derivative. IC50 for the native toxin was 0.13 nM (Table I). Ion-exchange chromatography of a sample treated with 1 M hydroxylamine (pH 7.0) gave a chromatogramme with the same number of peaks as in Fig. 1, nor was any peak signifi-
Localization of acetyl groups Acetyl derivatives were reduced with dithiothreitol and alkylated with 4-vinylpyridine as described by Hermodson et al. [7]. The resulting pyridylethyl derivatives were desalted by HPLC on a C4 reversed phase column using a gradient of 0.1% trifluoroacetic acid versus 0.08% trifluoroacetic acid in 60% acetonitrile and freeze-dried. The pyridylethyl derivative containing the most reactive amino group (the largest peak in ion-exchange chromatography of the acetyl derivatives) was digested with V8-protease: 0.1 M ammonium bicarbonate (pH 7.9), molar ratio e n z y m e / s u b s t r a t e 1/25, 37°C and 2 h. The peptide mixture was analyzed by plasma desorption mass spectrometry [8] with a Bioion 20 mass spectrometer (Applied Biosystems AB, Uppsala, Sweden). The peptides were isolated on a C18 reversed phase column in the same system as above. The peptide containing the acetyl group was submitted to sequence determination by Edman degradation with an automatic gas phase sequencer (Applied Biosystems Protein Sequencer 470A) and on line analysis of the phenylthiohydantoin derivatives with an Applied Biosystems 120A P T H Analyzer.
0.50-
G25r,.)
IZ
F
0
~00
i
i
!
|
|
0
20
40
60
80
RETENTION
TIME
i
100 (rain)
Fig. 1. Isolation of acetylated derivatives of fasciculin 2 by H P L C ion-exchange chromatography on a TSK-SP-5PW column ( 7 . 5 x 7 5 mm) equilibrated with water. Freeze-dried sample was dissolved in H 2 0 and applied to the column. After sample application the column was eluted for 4 rain with H 2 0 and then for 80 min with a linear gradient of H 2 0 vs. 1 M a m m o n i u m acetate (pH 6.8). The concentration of acetate increased by 2.5 m M (0.25%) per min. Flow rate 0.7 m l / m i n .
TABLE I
Acetyl derivatives of fasciculin 2 Derivative (numbering as in Fig. 1)
Acetyl groups per molecule
Remaining activity a %
Acetyl in position
A B C D E F G H I J
2b 2 2 2 1 1 1c 1 1 0
8 15 11 21 28 40 43 35 72 100
not determined Lys 58, Lys 25? not determined Lys 25, N-terminus Lys 32 Lys 58 Lys 51 Lys 25 N-terminus native toxin
a IC50 for native toxin = 0.13 nM. Remaining activity = IC50 for native t o x i n / I C s o for derivative. b Determined only from ratio cpm/A276. c Determined only by mass spectrometry.
cantly smaller. This indicates that tyrosine residues were not acetylated.
Localization of the acetyl groups 1. Derivative Ho The pyridylethyl derivative was digested with V8 protease that cleaves with high specificity at the C-terminal side of Glu residues. Cleavage should occur after GIu 19 and Glu 48 (Fig. 2). Three fragments with the masses 1675, 2297 and 3702 Da were detected by mass spectrometry. The two smallest fragments were the peptides 1-19 and 50-61, since the masses obtained by mass spectrometry agree very well with the composition data, 1676 and 2299 Da, respectively. The largest fragment had a mass 43 Da higher than for the peptide 20-48 and contained evidently an
+
40
-
_
2. The primary structure of fasciculin 2 (20). The toxin has an unusual distribution of charges. T h e second loop is very cationic and the third loop anionic. Fig.
acetyl group (42 Da) either on the side chain of Lys 25 or Lys 32. The peptide 20-48 was isolated by reversed phase chromatography on a C4 column and submitted to Edman degradation. The following sequence was obtained: Asn-Ser-Cys-Tyr-Arg-?-Ser-Arg-Arg-His-ProPro-Lys-Met-Val-Leu-Gly-Arg. To confirm that position 6 contained an acetylated Lys, a peptide Ala-AlaLys-Leu was synthesized by the Merrifield method and the side chain of Lys was reacted with acetic anhydride before the cleavage of the peptide from the resin. The phenylthiohydantoin derivative at step 3 in the Edman degradation eluted with the same retention time as the phenylthiohydantoin derivative at position 6 of the fragment 20-48. Acetylation of Lys 25 decreased the activity by 65% (Table 1). Circular dichroism spectra of derivative H and native fasciculin 2 were identical indicating that acetylation did not have a large effect on the conformation. 2. Derivatives E, F, G and I. To confirm the specificity of lysylendopeptidase, the pyridylethyl-derivative of H was digested with this enzyme and peptides were isolated by reversed phase chromatography. The peptides 1-32, 33-51 and 52-58 were detected by mass spectrometry but not the peptide 25-32 or the C-terminal tripeptide. The enzyme split as expected after Lys residues but not if the side chain was acetylated. The sample for mass spectrometry is applied on a filter that is washed to remove excess salt and during that process the C-terminal tripeptide was probably lost. The pyridylethyl derivatives of E, F, G and I were digested with lysyl endopeptidase, the peptides were isolated and their molecular weights were determined by mass spectrometry. The relative amounts of monoacetyl derivatives reflect the reactivity of the amino groups. Acetyl-Lys 25 accounted for 75%, acetyl-Lys 58 for 14% and the derivatives of the N-terminus, Lys 32 and 51 each for 3-4%. 3. Derivatives B and D. The diacetyl derivatives B and E were treated in the same way. Derivative B had the acetyl groups on Lys 58 and probably on Lys 25 and D on the N-terminus and Lys 25 (Table I). Discussion The binding between fasciculin and acetylcholinesterase is strong as indicated by a Ki of about 10 -H M [4-6]. This should result from interaction of several amino acid residues in the toxin with the enzyme. A modification of one of these residues should not abolish but significantly decrease the activity [9,10]. But the decrease in activity can also depend on structural perturbations caused by the modification.
Acetylation of the four e-amino groups reduced the activity between 57% (Lys 51) and 72% (Lys 32) while acetylation of the a-amino group had less effect and decreased the activity only by 28%. But as stated above the drop in activity can have two reasons. It is likely, however, that the lower activity of some of the monacetyl derivatives depends on modification of an amino group in the active site and for others on structural perturbations. Acetylation of Lys 25 reduced the activity by 65% without producing any structural changes, since the circular dichroism spectrum was not affected. Therefore, Lys 25 is probably in the active site. Indirect evidence for the role of Lys 32 and 51 can be obtained from a comparison of sequences of homologous toxins and from the three-dimensional structure of fasciculin [10]. A large number of snake toxins are homologous to fasciculin, such as a-neurotoxins which bind to the nicotinic acetylcholine receptor and cardiotoxins that increase the permeability of biological membranes. At the present, sequences of 175 homologues are known [11], A homology alignment with Cys-residues in the same positions shows that Lys 32 is unique for fasciculins and it should therefore be of special importance. Lysine in positions corresponding to 25, 51 and 58 are found frequently in many toxins. However, an amino acid that is located in the same position in different types of toxins can still be part of the active site. This has been shown for oxiana II, a short a-neurotoxin from Naja naja oxiana. By using spin labeled, fluorescent and photoaffinity derivatives it was possible to show that this lysine residue interacts with the acetylcholine receptor [12,13]. The three dimensional structure of fasciculin 1 has been determined recently [10]. The two fasciculins differ only at position 47 where fasciculin 1 has tyrosine and fasciculin 2 asparagine. Fasciculin 1 has a similar overall structure as short a-neurotoxins [14,15] and cardiotoxins [16], a dense core with disulphide bridges and the three loops Cys 3 to Cys 22, Cys 22 to Cys 39 and Cys 52 long and extended as the central fingers of a hand. Lys 25 is on the opposite side of the toxin molecule compared Lys 32 and 51 which form a cationic quartet together with Arg 24 and 28. The side chain of Lys 58 points in 90 ° angle from that of Lys 32 (Fig. 3). The cluster of cationic groups probably has a functional role as already suggested [10]. It contains Lys 32 which is unique for fasciculin. Acetylation of both lysine residues had a large effect decreasing the activity by 72% (Lys 32) and 57% (Lys 51). Thus, indirect evidence from chemical modification and structural data suggest that Lys 32 and 51 have a functional role. The evidence for Lys 25 is based on chemical modification and circular dichroism spec-
K25
K32 K51 Fig. 3. Schematic ribbon drawing of the three-dimensional structure of fasciculin 1, with E-structures indicated by arrows. The molecule has no a-helices. The side chain of Lys 58 points into the plane of the figure and the side chains of the other lysine residues are approximately in the plane of the figure. The data programme MOLSCRIPT (21) was used for drawing the figure.
troscopy. Functionally important amino acids should be then located on both sides of the fasciculin molecule (Fig. 3). Acetylation of the a-amino group and Lys 58 was also followed by loss of activity, but structural data cannot give any further indication of their possible functional role. Lys 25 has the most reactive amino group, acetyl-Lys 25 accounts for 75% of the monoacetyl derivatives. The reactivity of an amino group depends on its pKa and accessibility [17] and these factors seem to combine in fasciculin to give Lys 25 its superreactivity. The corresponding Lys in other toxins can also be very reactive as in the short a-neurotoxin erabutoxin b where it has the most reactive amino group [18]. The three-dimensional structure of AChE from Torpedo californica has been determined [19]. The active site is at the bottom of a deep gorge and the peripheral anionic site with the binding site for propidium [16] at the rim of the gorge. Since fasciculin and propidium have a common binding site [5], the toxin should also bind to the rim.
Acknowledgements This work was supported by the Swedish National Science Research Council, by O.E. och Edla Johanssons Stiftelse, by the International Program in the Chemical Sciences (IPICS), Uppsala University, and by the International Foundation for Science (IFS), Stockholm. We thank Dr. Andr~ M6nez for recording of the circular dichroism spectra, Dr. Juan Fontecilla-Camps for information on the three-dimensional structure of
fasciculin and Dr. Bengt Westerlund for computer drawing of the structure. References 1 Rodrfguez-Ithurralde, D., Silveira, R., Barbeito, L. and Dajas, F. (1983) Neurochem, Int. 5, 267-274. 2 Karlsson, E., Mbugua, P.M. and Rodr~guez-Ithurralde, D. (1985) Pharmac. Ther. 30, 259-276. 3 Cervefiansky, C., Dajas, F., Harvey, A.L. and Karlsson, E. (1991) In: International Encyclopedia of Pharmacology and Therapeutics: Snake Toxins (Harvey, A.L., ed.), pp. 131-164, Pergamon Press, New York. 4 Puu, G. and Koch, M. (1990) Biochem. Pharmacol. 40, 2209-2214. 5 Riordan, J.F. and Vallee, B.L. (1967) Methods Enzymol. 11, 570-576. 6 Karlsson, E., Mbugua, P.M. and Rodr~guez-lthurralde, D. (1984) J. Physiol. Paris 79, 232-240. 7 Hermodson, M.A., Ericsson, L.H., Neurath, H. and Walsh, K.A. (1973) Biochemistry 12, 3146-3153. 8 Sundqvist, B. and MacFarlane, R.D. (1985) Mass Spectrom. Rev. 4, 421-460. 9 Karlsson, E. and Sundelin, J. (1976) Toxicon 14, 295-306.
10 Le Du, M.H., Marchot, P., Bougis, P. and FonteciUa-Camps, J. (1992) J. Biol. Chem. 267, 22122-22130. 11 Mebs, D. and Claus, I. (1991) In: International Encyclopedia of Pharmacology and Therapeutics. Snake Toxins (Harvey, A.L., ed.), pp. 425-447. 12 Tsetlin, V.I., Karlsson, E., Arseniev, A.S., Utkin, Yu.N., Surin, A.M., Pashkov, V.S., Pluzhnikov, K.A., Ivanov, V.T., Bystrov, V.F. and Ovchinnikov, Yu. A. (1979) FEBS Lett. 106, 47-52. 13 Kreienkamp, H.J., Utkin, Yu.N., Weise, C., Machold, J., Tsetlin, V.I. and Hucho, F. (1992) Biochemistry 31, 8239-8244. 14 Tsernoglu, D. and Petsko, G.A. (1976) FEBS Lett. 68, 1-4. 15 Low, B., Preston, H.S., Sato, A., Rosen, L.S., Searl, J.E., Rudko, A.D. and Richardson, J.S. (1976) Proc. Natl. Acad. Sci. USA. 73, 2991-2994. 16 Smith, J.L., Corfield, P.W.R., Hendrickson, W.A. and Low, B. (1988) Acta Cryst. A44, 357-368. 17 Bosshard, H.R. (1979) Methods Biochem. Anal. 25, 273-301. 18 Hori, H. and Tamiya, N. (1976) Biochem. J. 153, 217-222. 19 Sussman, J.l., Harel, M., Frolow, F., Oefner, C., Goldman, A., Toker, L. and Silman, I. (1991) Science 253, 872-879. 20 Viljoen, C.C. and Botes, D.P. (1973) J. Biol. Chem. 248, 49154919. 21 Kraulis, P. (1991) J. Appl. Crystallogr. 24, 946-950.