- Email: [email protected]
Noxiustoxin 2, a novel K+ channel blocking peptide from the venom of the scorpion centruroides noxius Hoffmann
Pergamon
PII: S0041-0101(96)00029-3
Tmcon, Vol. 34, No. 8, pp. 913-922, 1996 Copyright 0 1996 Else&x Science Ltd Printed in Great Britain. All rights reserved 004-0101/96 $15.00 + 0.00
NOXIUSTOXIN 2, A NOVEL K+ CHANNEL BLOCKING PEPTIDE FROM THE VENOM OF THE SCORPION CENTRUROIDES NOXIUS HOFFMANN ALEJANDRO
R. NIETO,’ GEORGINA B. GURROLA,’ and LOURIVAL D. POSSANI’*
LUIS VACA’,2
‘Department of Molecular Recognition and Structural Biology Institute of Biotechnology, Universidad National Aut6noma de Mexico Avenida Universidad, 2001, Apartado Postal 510-3, Cuemavaca 62271, Mexico and *Department of Bioenergetics, Institute of Cellular Physiology, Universidad National Autonoma de Mexico, Ciudad Universitaria, Mexico 04510, Mexico (Received 20 November 1995; accepted 31 January 1996)
A. R. Nieto, G. B. Gurrola, L. Vaca and L. D. Possani. Noxiustoxin 2, a novel K+ channel blocking peptide from the venom of the scorpion Centruroides noxius Hoffmann. Toxicon 34, 913-922, 1996.-A novel peptide called Noxiustoxin 2 (NTX2) was purified from the venom of the scorpion Centruroides noxius and characterized chemically and functionally. It is composed of 38 amino acid residues linked by three disulfide bridges and its primary structure is 61% identical to that of Noxiustoxin (NTX). It is not toxic to mice (using up to 200 fig/20 g mouse weight) and crustaceans (up to 30 pg/g of crayfish), but has a paralysing effect on crickets (30 pg/g animal). It displaces the binding of [‘251]NTXto rat brain synaptosome membranes with a Ki of 0.1 PM, in comparison NTX has a K, of 100 pM. Similarly, using single Ca2+ activated K+ channels of small conductance obtained from cultured bovine aortic endothelial cells it was shown that NTX2 is over two logarithm units less potent than NTX in producing 50% blockade of the probability of opening the channels. NTX2 is not recognized by a panel of six distinct monoclonal antibodies against NTX, however it is recognized by polyclonal antibodies raised in mouse, with native NTX. Primary structure comparison of both NTX and NTX2 suggests that the N-terminal segments of these peptides are important for channel affinity. Copyright 0 1996 Elsevier Science Ltd
INTRODUCTION
Scorpion toxins have been instrumental in defining both structural and mechanistic properties of ion channels, and have been used with other blocking agents to examine in detail the various conductances normally present in several types of cells (Hille, 1992). Most scorpion toxins against K’ channels (K’ toxins) are short peptides, composed of 31-39 amino acid residues, with the same general three-dimensional structure: a short *Author to whom correspondence
should be addressed. 913
914
A. R. NIETO
et al.
alpha helix and three strands of beta sheet structures, stabilized by three disulfide bridges (Menez et al., 1992). However, their primary structures show a highly variable amino acid sequence (Garcia et al., 1995; Martin et al., 1994; Menez et al., 1992; Possani, 1984), that is thought to be responsible for the differential recognition and/or affinity of these peptides towards distinct sub-types of K+ channels. The affinity differences can span over six orders of magnitude, depending on the sub-type of K+ channels under assay. For example Noxiustoxin (NTX), the first K + -toxin ever described (Carbone et al., 1982; Possani et al., 1982), binds with 100 pM affinity to the K+ channels present in rat brain synaptosomes (Gurrola and Possani, 1995), whereas it binds with 450 nM affinity to Ca’ + -dependent K + channels of rabbit skeletal muscle (Valdivia et al., 1988), and with 310 nM to Ca2 + -dependent K + channels of epithelial cells (Vaca et al., 1993). Charybdotoxin (ChTX), purified from Leiurus quinquestriutus venom, is a homologous peptide to NTX (Valdivia et al., 1988), with a 46% identity in primary structure (Gurrola and Possani, 1995), has a iyd of 25-30 pM to plasma K+ channels of membranes prepared from rat brain synaptosomes (Vazquez et al., 1990) and 1.8 nM to Ca*+-dependent KC channels from rabbit skeletal muscle (Valdivia et al., 1988). Similarly, Margatoxin, another K +-toxin isolated from Centruroides margarita&s, binds to purified rat brain synaptic plasma membrane vesicles with a dissociation constant of 0.1 pM, under equilibrium binding conditions (Knaus et al., 1995). It seems possible that nature has evolved in parallel different types of K+ toxins for different types (or sub-types) of K+-channels. If we want to corroborate this hypothesis and to correlate the various structures of K+-toxins with the modulation of activity of various K+-channels, it is necessary either to produce artificial mutants, by site directed mutagenesis, as done for ChTX (Park and Miller, 1992) or to find the distinct mutants already selected by nature, isolating similar peptides from various sources, and similar K + -channels also from various tissues. In this communication we describe the isolation and complete amino acid sequence determination of a NTX-like peptide, called NTX2. The new peptide was assayed in binding and displacement experiments, using brain synaptosome membranes, and Ca2 + -dependent K + channels of epithelial cells. Several structural features of both peptides were analyzed and are discussed in view of their binding properties to a panel of monoclonal antibodies and to their affinities towards the K+ -channels assayed. The N-terminal segment of these peptides seems to be important for channel affinity. MATERIALS
AND
METHODS
Source of venom Scorpions of the species C. noxius maintained alive in the laboratory were monthly anesthetized with CO, and milked for venom, by means of electrical stimulation. The venom was placed in double distilled water and centrifuged at 15,OOOg for 15 min. The supernatant was freeze-dried and stored at - 20°C until use.
Separation procedure The soluble venom was initially fractionated in a Sephadex G-50 column as described (Possani et al., 1982). Fraction II from this column containing peptides toxic to mice was subsequently fractionated by ion exchange column on carboxymethyl-cellulose (CM-cellulose) resins, also as described (Possani et al., 1982). The fraction containing NTX-like peptides was additionally separated by high performance liquid chromatography (HPLC), using a Cl8 reverse-phase column (Vydac, Hysperia, CA), of a Waters 600E chromatographer. equipped with a variable wavelength detector, and a WIPS 712 automatic sample injector. The final pure peptide was obtained by means of a step-gradient HPLC. Homogeneity of sample was verified by HPLC profile and by direct Edman degradation using an automatic sequencer (Martin et al., 1994). Amino acid analysis of peptides confirmed the molecular mass, purity of sample, and the amino acid sequence found.
Noxiustoxin 2, a Novel K+ Channel Blocking Peptide
915
During purification procedures this peptide was identified by an enzyme-linked-immune-assay (ELISA) technique using mouse antibodies, raised against NTX. These antibodies were obtained from two female mice of the strain Balb/c immunized repeatedly with 20 pg of NTX, per animal, with 15 days interval each time. The first injection (always s.c.) was in the presence of complete Freund’s adjuvant, and the subsequent ones with incomplete Freund’s adjuvant. The antibodies were prepared from blood collected after 7 weeks of immunization. Similar peptides were identified by the same ELISA method, but with a panel of six different monoclonal antibodies prepared as previously reported (H&ion et al., 1995). Lethality tests
A limited number of lethality tests were conducted with mice, crustaceans and crickets. Two albino mice (strains CDl) were injected intraperitoneally with 200 ng of NTX2 per 20 g mouse. None of the animals showed any symptoms of intoxication. Similarly, in our crustacean assay (sweet water shrimp), 30 pg of NTX per g of body weight of crustacean did not affect the animals. However, adult crickets injected with 30 pg of NTX2 showed symptoms of intoxication: paralysis during the first 15 min, followed by a clear hind limb paralysis and uncoordinated movements, but recovered after a couple of hours. Further lethality tests were prevented due to the limited amounts of highly purified NTX2 Amino acid analysis and sequencing
Samples (about 1 nmole each) of the pure native NTX2, and their fragments generated by enzymatic digestion, were analyzed in a Beckman 6300E amino acid analyzer, after acid hydrolysis for 20 hr in 6 N HCI at I1 0°C. An ahquot of pure NTX2 (100 fig) was reduced and alkylated with iodoacetic acid, as described (Martin et af., 1994). An automatic ProSequencer (Millipore model 6400/6600) was used in order to determine the amino acid sequence of 1. native NTX2; 2. its reduced and carboxy-methylated derivative (RC-toxin); and 3. their HPLC-purified fragments produced by trypsin hydrolysis and cyanogen bromide cleavage Two types of polyvinylidene difluoride membranes derivatized with aryl-amine (AM-membranes) and 1,4-phenylene diisothyocyanate (DITC-membranes) were used to attach the peptides for sequencing, as recommended by the manufacturer (Millipore Co., Milford, MA, U.S.A.). Trypsin from Boehringer (Mannheim, Germany) was used at I:50 ratio dilution (trypsin:NTXZ) in the presence of 20 mM Tris-HCI buffer pH 7.8, for 12 hr, at room temperature, while cyanogen bromide (I mg/ml) cleavage of RC-toxin (30 pg) was obtained by incubation in 0.4 ml of 70% formic acid, at room temperature, overnight. All peptides were purified by HPLC prior to sequence determination. Protein content
The concentration of peptides used for binding and electrophysiological on amino acid analysis.
measurements were estimated based
Binding and displacement assays
NTX was radiolabeled with [125]iodineand used for binding and displacement assays, as described (Valdivia membranes (fraction P3) were prepared by the method of Catterall et al. (1979), and used for assessing the capability of both NTX and NTX2 to displace the binding of [‘Z51]NTX (Valdivia et al., 1992).
et al., 1992). Rat brain synaptosomal
Electrophysiological experiments
Bovine aortic endothelial cells were cultured and assayed as described (Vaca et al., 1993). The outside-out configuration of the patch clamp (Hamili et a/., 1981) was used to study single channels obtained from excised patches from single endothelial cells. The recordings and single channel analysis (PO= open time/total time) were performed also, as described (Vaca et al., 1993).
RESULTS AND DISCUSSION
Peptide pur$cation
and amino acid sequence
A novel peptide was purified from the venom of the scorpion C. noxius by essentially the same procedure originally described (Possani et al., 1982), with the addition of two
916
A. R. NIETO et al.
HPLC steps. Briefly, the soluble venom was separated by gel-filtration chromatography in Sephadex G-50 [data identical to Possani et al. (1982)]. Fraction II from this column was applied to a CM-cellulose column providing the separation shown in Fig. l(a). Sub-fraction 1lb, which contains the K+-toxins as indicated by positive ELISA with anti-NTX antibodies, was further selected for additional separation on HPLC as shown
Volume 1.c
( ml 1
b iQS-
a LI OO
/ /
./I 0
/
/ //
50
mm o\”
'12
//
, 20
I
20 4Q,’ Time (/min) / / /
3 L
40
)
Time (min) Fig. 1. Ion-exchange separation of fraction 11. (a) A CM-Cellulose column (0.9 x 30 cm) was loaded with 100 mg of fraction II from Sephadex G-50 separation of C. noxius venom, in the presence of 20 mM ammonium acetate buffer, pH 4.7 and eluted with a linear gradient from 0 to 0.5 M of NaCl in this buffer. The flow rate was 30 ml/hr and tube-fractions containing 2.5 ml were collected. (b) Sub-fraction 11b was further applied to a Cl8 reverse-phase column of a HPLC apparatus and separated with a linear gradient from solution A (0.12% trifluoroacetic acid in water) to 60% solution B (0.10 trifluoacetic acid in acetonitrile). Component 12 was finally applied to a C4 reverse-phase column [inset Fig. l(b)] and eluted with a constant solvent concentration of 40% solution B.
Noxiustoxin 2, a Novel K+ Channel Blocking Peptide
917
in Fig. 1(b). The first HPLC separated about 20 components, the largest peak corresponds to NTX, and component 12 with an elution time of 36.25 min contained NTX2. Since this chromatogram showed some overlapping traces, component 12 was re-applied into a C4 reverse-phase analytical column, using a step gradient of 40% acetonitrile in water, containing 0.1% trifluoroacetic acid [see inset Fig. l(b)]. A couple of small contaminants were eliminated, and the main component was pure toxin NTX2, which was used for chemical and physiological characterization. Final recovery of NTX2 shows that it is 0.13% of the soluble venom. The identification of the NTX-like peptides was performed with polyclononal antibodies. Figure 2(a) shows the positive HPLC fractions recognized in the ELBA test. Peptides eluting at eight distinct retention times gave a positive result (numbers 2, 5, 6, 8, 9, 10, 12, and 15), using polyclonal anti-NTX antibodies. A little < 50% recognition was obtained with component 12, when compared to the control [NTX in Fig. 2(a)]. After
12345
NTX
67
NTX-2
0
’
3 10 11 12 13 14 15 16 NTX
NTX
NTX-2
Fig. 2. ELISA recognition of fractions from HPLC. (a) Percentage of positive recognition of sub-fractions from Fig. l(b), by polyvalent antibodies against NTX, using an ELISA format (Martin et al., 1994). Fractions correspond to the following elution times (in minutes shown in parentheses): 1 (23.90); 2 (24.74); 3 (25.19); 4 (25.26); 5 (28.32); 6 (29.76); 7 (31.04); 8 (31.85); 9 (33.26); 10 (34.62); 11 (35.63); 12 (36.26); 13 (36.77); 14 (37.92); 15 (38.11); 16 (39.90). The last is NTX used as control, and fraction number 12 corresponds to NTX2. (b) Similar results obtained with NTX (control) and NTX2 after fully purified by step-gradient HPLC [inset Fig. l(b)]. (c) Similar experiments using a set of six different monoclonal antibodies against NTX (left side): clones BNTX4, - 12, - 14, - 16, - 18, and - 21 from H&ion er al. (1995) and the same monoclonals (right side) against NTX2 (in that order, from left to right side).
918
A. R. NIETO et al.
1
10
TIINEKCFAT
:------d-m :.T2.>:... :======r== :e l =*=.=*=
20 30 38 SQCWTPCKKA IGSLQSKCMN GKCKCYNG ---me> :- .-.cb-> Tl . . ...>.. ..T4..> :.T3> st===dltc= ------------__---, .=.=.am=,= .=*=*=*=*3 .=.> aa
Fig. 3. Primary structure determination of NTX2. Numbers on top indicate the amino acid position in the sequence (one letter code). Underlined are the sequences obtained with direct automatic Edman degradation (-d-), residues l-16; cyanogen bromide cleaved peptide (.-cb.-), residues 29-37; tryptic fragments (.Tl., eluted at 32.50 min on the HPLC; .T2., at 23.74; .T3., at 26.46; and .T4., at 24.43 min) corresponding to residues from 7 to 18, 1 to 6, 33 to 37, and 18 to 27, respectively; RC-toxin attached to DITC membranes ( = ditc = ), residues l-30; and, finally RC-toxin attached to aryl-amine membranes ( = .am = .), residues l-34. Amino acid determined by amino acid analysis is indicated by aa.
the second HPLC separation, inset Fig. l(b), the pure NTX2 was still positive in the ELISA assay, as shown on Fig. 2(b), while the same homogeneous toxin was not recognized by a panel of six different monoclonal antibodies raised against NTX [Fig. 2(c)]. Figure 3 shows the results of the amino acid sequence of NTX2. Direct sequencing of native toxin permitted the identification of the first 15 residues, while the analysis of RC-toxin attached to aryl-amine- and DITC-membranes provided identification of up to residue number 35. Cyanogen bromide cleavage was important to confirm the C-terminal fragment, from Met29 to Asn37. Several overlapping fragments were obtained by tryptic digestion of RC-toxin, as indicated in Fig. 3. The last residue was surmised based on the results of the amino acid analysis of both the native NTX2 and the C-terminal fragment produced by cyanogen bromide cleavage. Thus, NTX2 is a 38 amino acid residue peptide containing six half-cystines. These results may suggest the presence of three disulfide bridges, as demonstrated for the other known K +-toxins (Possani et al., 1982; Menez et al., 1992; Garcia et al., 1995). Recognition of K+ channels Since NTX2 was 61% identical to NTX, it was interesting to determine the ability of NTX2 to recognize KC-channels. Two preparations were used, as shown in Fig. 4. By means of binding and displacement experiments, using radiolabeled NTX (Valdivia et al., 1992) it was found that both native NTX and NTX2 were effective in competing for similar binding sites (K+ channels) of membranes obtained from rat brain synaptosomes [Fig. 4(a)]. Interestingly, while cold NTX displaced radiolabeled-NTX with a K, close to 100 pM, NTX2 was able to displace 50% binding only at 100 nM. In other words, NTX2 has a thousandfold less affinity than NTX, for the KC-channels of synaptosomes. A second preparation, the outside-out configuration of the patch-clamp method [Fig. 4(b)], was used to study the effect of NTX and NTX2 on single Ca2+-activated K+ channels of small conductance obtained from cultured bovine aortic endothelial cells (Vaca et al., 1993). In this preparation, the 50% probability of opening (P,,) of these channels was decreased by at least two orders of magnitude when comparing NTX and NTX2, the second being less effective than the first one. Structure-function relationship Why would the scorpion venom contain more than one such toxin? And why are their affinities so different? If the known K+ -toxins are compared (Table 1) it is evident that
Noxiustoxin
2, a Novel
0-
K’
Channel
-12-11-10-9
Blocking
-6 -7-6
Log [peptide]
Peptide
919
-5
M
1.0
b
0.0
PO 0.4 0.2 0.6 ~ 0.0
-10-6-6-4-2
0
Log [peptide]
M
Fig. 4. Binding properties and channel recognition of NTX2. (a) Binding and displacement experiments were performed using brain synaptosomal membranes (85 ng per assayed-point) incubated with [‘2’I]Noxiustoxin (50 pM, specific activity 2100 Ci/mmole) and displaced with increasing concentration of cold NTX (0) and NTX2 (0). Mean value of triplicated measurements are plotted. (b) Concentration-response curve for NTX (0) and NTX2 (0) that affected the obtained probability of opening (PJ of single Ca2 + activated K + channels of small conductance from cultured bovine aortic endothelial cells. Mean value of quadruphcated measurements are shown.
there is an important variation at the level of their primary structures, which range from 26 to 61% identity to NTX2. The American species from the genus Centruroides and Tityus have about 50-61% identity, while the scorpions from the Old World of the genus Leiurus, Androctonus and Buthus have only 2644% identity with that of NTX2. However, the most important point is that out of these 40 positions shown in Table 1 ( - 1 position was included in order to increase similarity) only the cysteines and two lysines, at positions 28 and 33 are absolutely conserved. The consensus sequence for all K + -toxins, for which the full sequences are known, presents several conserved residues at the carboxyl-terminal segment. This suggests that the C-terminal segment is important either for function or structural reasons, positioning the N- and C-terminal regions in the right conformation for channel recognition. From the three-dimensional NMR-solution of various K+-channel toxins it is known that the N-terminal and C-terminal region are relatively close, maintained by the three disulfide bridges (Bontems et al., 1992; Meunier et al., 1993; Fernandez et al., 1994; Johnson et al., 1994; Krezel et al., 1995). It is also known, at least for the case of ChTX (Park and Miller, 1992) that the Lys27, equivalent to position 28 of Table 1, is a ‘crucial residue’ for the interaction with the Shaker B K + -channels (Stampe
A. R. NIETO et al.
920
Table 1. Comparative amino acid sequence of K+ channel toxins Toxin 1 NTX2 NTX MgTX CllTX 1 TsKa KlTX AgTXl Lq2 AgTX2 AgTX3 IbTX ChTX LeTX I PO5 Consensus
Identity (%)
Amino acid seauence -TIINEKCFAT -TIINVKCTSP -TIINVKCTSP -1TINVKCTSP -VFINAKCRGS GVEINVKCSGS GVPINVKCTGS pEFTQESCTAS GVPINVSCTGS GVPINVPCTGS pEFTDVDCSVS pEFTNVSCTTS -AF----C-NL TV--NL
--
10
20
30
SQCWTPCKKA KQCSKPCKEL KQCLPPCKAQ QQCLRPCKDR PECLPKCKEA PQCLKPCKDA PQCLKPCKDA NQCWSICKRL PQCIKPCKDA PQCIKPCKDA KECWSVCKDL KECWSVCQRL RMCQLSCRSL RRCQLSCRSL -Cd-
IGSLQS-KCM YGSSAGAKCM FGQSAGAKCM FGQHAGGKCI IGKAAG-KCM GMRF+KCM GMRF-G-KC1 HNTNRG-KCM GMRF-G-KCM GMRF-G-KCM FGVDRG-KCM HNTSRG-KCM -GL-LG-KC1 -GL-LG-KC1 KC-
NGKCKCYNG NGKCKCYNN NGKCKCYPH NGKCKCYPNGKCKCYPNRKCHCTPK NGKCHCTPK NKKCRCYSNRKCHCTPK NRKCHCTPK GKKCRCYQNKKCRCYSGDKCECVKH GVKCECVKH -KC-C-
39 100 61 56 51 50 44 44 42 41 41 37 34 26 26
NTXZ, from this work; CllTX 1, toxin 1 from Martin et al. (1994); NTX, noxiustoxin from Possani et al. (1982); MgTX (margatoxin), ChTX (charybdotoxin), IbTX (iberiotoxin), Lq2 (L. quinquestriatus toxin 2) and AgTXl to AgTX3 from L. quinquestriatus var. hebraeus [reviewed by Garcia et al. (1995)]; LeTX I, leirutoxin 1 (Auguste et al., 1990; Chicchi et al., 1988). KlTX, kaliotoxin (Crest et al., 1992); TsKa from the venom of T. serrulatus (Rogowski et al., 1992), and PO5 from the venom of A. mauretanicus mauretanicus (Sabatier et al., 1993). Consensus stands for positions in which all the amino acids are conserved in all sequences. Gaps (-) were introduced to enhance similarities.
et al., 1994). However, for NTX the N-terminal amino acid sequence is very important. Actually, a nonapeptide corresponding to this section is toxic to mice (Gurrola et al., 1989) and recognizes some sub-types of K+ channels (Vaca et al., 1993). NTX2 presents two highly similar amino acid sequences to NTX, one being the N-terminal (first eight residues, except for position 5) and the other at the C-terminal segment (positions 28-38). The presence of a negatively charged amino acid (Glu) in position 5, substituting for the hydrophobic valine in NTX, probably renders this segment unfavorable for binding to the K + channels, as shown in Fig. 4. Thus, functional differences shown to exist between NTX and NTX2 might reside preferentially, either at the important changed glutamic acid of position 5, or at some other highly variable region of the internal segment. However, this does not rule out the possibility that both the N-terminal and the internal segments (corresponding to positions 9-27), are important for the functional differences observed. Furthermore, worth mentioning is the lack of glycine at position 27 which was claimed to be essential for proper folding of ChTX (Bontems et al., 1992; MCnez et al., 1992). This residue is present in all the other K+-toxins, but is missing in NTX2. In conclusion, it is quite possible that all the variations we observed in terms of primary structure, despite the fact of a common structural motif at the three-dimensional level, (Menez et al., 1992) are necessary modifications evolved by scorpions to cope with the co-evolving types and sub-types of K+ -channels. This might answer the question why scorpions have so many different toxins in the same venom. Additionally, we should consider the possibility of species specificity discussed earlier. Some of these peptides are designed to recognize different tissues from different taxonomic groups, like the case of NTX2, which is toxic only to insects.
Noxiustoxin 2, a Novel K+ Channei Blocking Peptide
921
Acknowledgements-The technical assistance of M.Sc. Fernando Z. Zamudio, Timoteo C. Olamendi-Portugal, and Sandra Contreras is greatly acknowledged. Supported in part by grants from Howard Hughes Medical Institute (No. 75191-527104), from National University of Mexico (DGAPA No. IN-205893 and IN-206994), National Council of Research and Technology of Mexico (CONACyT No.4734N), and European Commission (CII*-CT 94-0045). A. Nieto received a scholarship from CONACyT.
REFERENCES Auguste, P., Hugues, M., Grave, B., Gesquitre, J. C., Maes, P., Tartar, A., Romey, G., Schweitz, H. and Lazdunski, M. (1990) Leiurotoxin I (Scyllatoxin), a peptide ligand for Ca’+ -activated K+ channels. /. hiol. Chem. 265, 47534759.
Bontems, F., Gilquin, B., Roumestand, C., Mtnez, A. and Toma, F. (1992) Analysis of side-chain organization on a refined model of charybdotoxin: structural and functional implications. Biochemistry 31, 77567764. Carbone, E., Wanke, E., Prestipino, G., Possani, L. D. and Maelicke, A. (1982) Selective blockage of voltage-dependent K’ channels by a novel scorpion toxin. Nature (London) 2%, 90-91. Catterall, W. A., Morrow, C. S. and Harfshorne, R. P. (1979) Neurotoxin binding to receptor sites associated with voltage-sensitive sodium channels in intact, lysed and detergent solubilized brain membranes. J. hiol. Chem. 254, 11379-11387. Chicchi, G. G., Gimenez-Gallego, G., Ber, E., Garcia, M. L., Winquist, R. and Cascieri, M. A. (1988) Purification and characterization of a unique, potent inhibitor of apamin binding from Leiurus quinquestriatus hebraeus venom. J. biol. Chem. 263, 10192-10197. Crest, M., Jacquet, G., Gola, M., Zerrouk, H., Benslimane, A., Rochat, H., Manuelle, P. and Martin-Eauclaire, M. F. (1992) Kaliotoxin, a novel peptidyl inhibitor of neuronal BK-type Ca*+-activated K+ channel characterized from Androctonus mauretanicus mauretanicus venom. J. biol. Chem. 267, 164&1647. Fernandez, I., Romi, R., Szendeffy, S., Martin-Eauclaire, M. F., Rochat, H., Van Rietschoten, J., Pons, M. and Giralt, E. (1994) Kaliotoxin (l-37) shows structural differences with related potassium cahnnel blockers. Biochemistry
33, 1425614263.
Garcia, M. L., Knaus, H. G., Munujos, P., Slaughter, R. and Kaczorowski, G. J. (1995) Charybdotoxin and its effects on potassium channels. Am. J. Physiol. (Cell Physiol) 269, Cl-ClO. Gurrola, G. B., Molinar-Rode, R., Sitges, M., Bayon, A. and Possani, L. D. (1989) Synthetic peptides corresponding to the sequence of Noxiustoxin indicate that the active site of this K + channel blocker is located on its amino-terminal portion. J. Neur. Transmission 77, 1l-20. Gurrola, G. B. and Possani, L. D. (1995) Structural and functional features of Noxiustoxin: a K” channel blocker. Biochem. Mol. Biol. Int. 37, 527-535. Hamill, 0. P., Marty, A., Neher, E., Sakamann, B. and Sigworth, F. J. (1981) Improved patch clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pfiigers Arch. 391, 855100. Herion, P., Gurrola-Briones, G., Sanchez, M. R., Saavedra, R. and Possani, L. D. (1995) Monoclonal antibodies against Noxiustoxin. Hybridoma 14, 247-251. Hille, B. (1992) Mechanisms of block. In: Ionic Channels of Excitable Membranes, pp. 39&422. (Hille, B., Ed.). Sunderland.: Sinauer Assoc. Inc. Johnson, B. A., Stevens, S. P. and Williamson, J. M. (1994) Determination of the three-dimensional structure of margatoxin by ‘H, “C, “N triple resonance nuclear magnetic resonance spectroscopy. Biochemistry 33, 15061-15070.
Knaus, H. G., Koch, R. O., Eberhart, A., Kaczorowski, G. J., Garcia, M. L. and Slaughter, R. S. (1995) [Y]Margatoxin, an extraordinarily high affinity ligand for voltage-gated potassium channels in mammalian brain. Biochemistry 34, 13627-13634. Krezel, A. M., Kasibhatla, C., Hidalgo, P., MacKinnon, R. and Wagner, G. (1995) Solution structure of the potassium channel inhibitor agitoxin 2: caliper for probing channel geometry. Protein Sri. 4, 1478-1489. Martin, B. M., Ramirez, A. N., Gurrola, G. B., Nobile, M., Prestipino, G. and Possani, L. D. (1994) Novel K + -channel blocking toxins from the venom of the scorpion Centruroides limpidus ltmpidus Karsch. Biochem. J. (Great Britain)
384, 5166.
Mtnez, A., Bontems, F., Roumestand, C., Gilquin, B. and Toma, F. (1992) Structural basis for functional diversity of animal toxins. Proc. R. Sot. Edinburgh 99B, 83-103. Meunier, S., Bernassau, J. M., Sabatier, J. M., Martin-Eauclaire, M. F., Van Rietschoten, J., Cambillau, C. and Darbon, H. (1993) Solution structure of POrNH2, a scorpion toxin analog with high affinity for the apamin-sensitive potassium channel. Biochemistry 32, 11969-l 1976. Park, C. S. and Miller, C. (1992) Mapping function to structure in a channel-blocking peptide: electrostatic mutants of charybdotoxin. Biochemistry 31, 7749-7755. Possani, L. D. (1984) Structure of Scorpion Toxins. In: Handbook of Natural Toxins, Vol. 2, pp. 513-550 (Tu, A. T. Ed.). New York: Marcel Dekker. Possani, L. D., Martin, B. M. and Svendsen, I. B. (1982) The primary structure of Noxiustoxin: a K’ Channel blocking peptide from the venom of the scorpion Centruroides noxius Hoffmann. Curlsberg Res. Commun. 47, 285-289.
922
A. R. NIETO et al
Rogowski, R. S., Collins, J. H. and Blaustein, M. P. (1992) Purification and structural characterization of a potent potassium channel blocker from scorpion Tityus serrulatus venom. Biophys. J. 61, A250 (abstract). Sabatier, J. M., Zerrouk, H., Darbon, H., Mabrouk, K., Benslimane, A., Rochat, H., Martin-Eauclaire, M. F. and Van Rietschoten, J. (1993) POS, a new Leiurotoxin I-like scorpion toxin: synthesis and structure-activity relationships of the alpha-amidated analog, a ligand of Ca’+ -activated K + channels with increased affinity. Biochemistry
32, 2163-2170.
Stampe, P., Kolmakova-Partensky, L. and Miller, C. (1994) Intimations of K+ channel structure from a complete functional map of the molecular surface of charybdotoxin. Biochemistry 33, 443450. Vaca, L., Gurrola, G. B., Possani, L. D. and Kunze, D. (1993) Blockade of a Kc. channel with synthetic peptides from Noxiustoxin: a K+ channel blocker. J. membr. Biol. 134, 123-129. Valdivia, H. H., Smith, J. S., Martin, B. M., Coronado, R. and Possani, L. D. (1988) Charybdotoxin and noxiustoxin, two homologous peptide inhibitors of the K + (Ca”) channel. FEBS Lett. 226, 280-284. Valdivia, H. H., Martin, B. M., Escobar, L. and Possani, L. D. (1992) Noxiustoxin and leiurutoxin III, two homologous peptide toxins with binding properties to synaptosomal membrane K + channels. Biochem. Int. 27, 953-962.
Vazquez, J., Feigenbaumm, P., King, V. F., Kaczorowski, G. J. and Garcia, M. L. (1990) Characterization of high affinity binding sites for charybdotoxin in synaptic plasma membranes from rat brain. J. biol. Chem. 265, 15564-15571.
PII: S0041-0101(96)00029-3
Tmcon, Vol. 34, No. 8, pp. 913-922, 1996 Copyright 0 1996 Else&x Science Ltd Printed in Great Britain. All rights reserved 004-0101/96 $15.00 + 0.00
NOXIUSTOXIN 2, A NOVEL K+ CHANNEL BLOCKING PEPTIDE FROM THE VENOM OF THE SCORPION CENTRUROIDES NOXIUS HOFFMANN ALEJANDRO
R. NIETO,’ GEORGINA B. GURROLA,’ and LOURIVAL D. POSSANI’*
LUIS VACA’,2
‘Department of Molecular Recognition and Structural Biology Institute of Biotechnology, Universidad National Aut6noma de Mexico Avenida Universidad, 2001, Apartado Postal 510-3, Cuemavaca 62271, Mexico and *Department of Bioenergetics, Institute of Cellular Physiology, Universidad National Autonoma de Mexico, Ciudad Universitaria, Mexico 04510, Mexico (Received 20 November 1995; accepted 31 January 1996)
A. R. Nieto, G. B. Gurrola, L. Vaca and L. D. Possani. Noxiustoxin 2, a novel K+ channel blocking peptide from the venom of the scorpion Centruroides noxius Hoffmann. Toxicon 34, 913-922, 1996.-A novel peptide called Noxiustoxin 2 (NTX2) was purified from the venom of the scorpion Centruroides noxius and characterized chemically and functionally. It is composed of 38 amino acid residues linked by three disulfide bridges and its primary structure is 61% identical to that of Noxiustoxin (NTX). It is not toxic to mice (using up to 200 fig/20 g mouse weight) and crustaceans (up to 30 pg/g of crayfish), but has a paralysing effect on crickets (30 pg/g animal). It displaces the binding of [‘251]NTXto rat brain synaptosome membranes with a Ki of 0.1 PM, in comparison NTX has a K, of 100 pM. Similarly, using single Ca2+ activated K+ channels of small conductance obtained from cultured bovine aortic endothelial cells it was shown that NTX2 is over two logarithm units less potent than NTX in producing 50% blockade of the probability of opening the channels. NTX2 is not recognized by a panel of six distinct monoclonal antibodies against NTX, however it is recognized by polyclonal antibodies raised in mouse, with native NTX. Primary structure comparison of both NTX and NTX2 suggests that the N-terminal segments of these peptides are important for channel affinity. Copyright 0 1996 Elsevier Science Ltd
INTRODUCTION
Scorpion toxins have been instrumental in defining both structural and mechanistic properties of ion channels, and have been used with other blocking agents to examine in detail the various conductances normally present in several types of cells (Hille, 1992). Most scorpion toxins against K’ channels (K’ toxins) are short peptides, composed of 31-39 amino acid residues, with the same general three-dimensional structure: a short *Author to whom correspondence
should be addressed. 913
914
A. R. NIETO
et al.
alpha helix and three strands of beta sheet structures, stabilized by three disulfide bridges (Menez et al., 1992). However, their primary structures show a highly variable amino acid sequence (Garcia et al., 1995; Martin et al., 1994; Menez et al., 1992; Possani, 1984), that is thought to be responsible for the differential recognition and/or affinity of these peptides towards distinct sub-types of K+ channels. The affinity differences can span over six orders of magnitude, depending on the sub-type of K+ channels under assay. For example Noxiustoxin (NTX), the first K + -toxin ever described (Carbone et al., 1982; Possani et al., 1982), binds with 100 pM affinity to the K+ channels present in rat brain synaptosomes (Gurrola and Possani, 1995), whereas it binds with 450 nM affinity to Ca’ + -dependent K + channels of rabbit skeletal muscle (Valdivia et al., 1988), and with 310 nM to Ca2 + -dependent K + channels of epithelial cells (Vaca et al., 1993). Charybdotoxin (ChTX), purified from Leiurus quinquestriutus venom, is a homologous peptide to NTX (Valdivia et al., 1988), with a 46% identity in primary structure (Gurrola and Possani, 1995), has a iyd of 25-30 pM to plasma K+ channels of membranes prepared from rat brain synaptosomes (Vazquez et al., 1990) and 1.8 nM to Ca*+-dependent KC channels from rabbit skeletal muscle (Valdivia et al., 1988). Similarly, Margatoxin, another K +-toxin isolated from Centruroides margarita&s, binds to purified rat brain synaptic plasma membrane vesicles with a dissociation constant of 0.1 pM, under equilibrium binding conditions (Knaus et al., 1995). It seems possible that nature has evolved in parallel different types of K+ toxins for different types (or sub-types) of K+-channels. If we want to corroborate this hypothesis and to correlate the various structures of K+-toxins with the modulation of activity of various K+-channels, it is necessary either to produce artificial mutants, by site directed mutagenesis, as done for ChTX (Park and Miller, 1992) or to find the distinct mutants already selected by nature, isolating similar peptides from various sources, and similar K + -channels also from various tissues. In this communication we describe the isolation and complete amino acid sequence determination of a NTX-like peptide, called NTX2. The new peptide was assayed in binding and displacement experiments, using brain synaptosome membranes, and Ca2 + -dependent K + channels of epithelial cells. Several structural features of both peptides were analyzed and are discussed in view of their binding properties to a panel of monoclonal antibodies and to their affinities towards the K+ -channels assayed. The N-terminal segment of these peptides seems to be important for channel affinity. MATERIALS
AND
METHODS
Source of venom Scorpions of the species C. noxius maintained alive in the laboratory were monthly anesthetized with CO, and milked for venom, by means of electrical stimulation. The venom was placed in double distilled water and centrifuged at 15,OOOg for 15 min. The supernatant was freeze-dried and stored at - 20°C until use.
Separation procedure The soluble venom was initially fractionated in a Sephadex G-50 column as described (Possani et al., 1982). Fraction II from this column containing peptides toxic to mice was subsequently fractionated by ion exchange column on carboxymethyl-cellulose (CM-cellulose) resins, also as described (Possani et al., 1982). The fraction containing NTX-like peptides was additionally separated by high performance liquid chromatography (HPLC), using a Cl8 reverse-phase column (Vydac, Hysperia, CA), of a Waters 600E chromatographer. equipped with a variable wavelength detector, and a WIPS 712 automatic sample injector. The final pure peptide was obtained by means of a step-gradient HPLC. Homogeneity of sample was verified by HPLC profile and by direct Edman degradation using an automatic sequencer (Martin et al., 1994). Amino acid analysis of peptides confirmed the molecular mass, purity of sample, and the amino acid sequence found.
Noxiustoxin 2, a Novel K+ Channel Blocking Peptide
915
During purification procedures this peptide was identified by an enzyme-linked-immune-assay (ELISA) technique using mouse antibodies, raised against NTX. These antibodies were obtained from two female mice of the strain Balb/c immunized repeatedly with 20 pg of NTX, per animal, with 15 days interval each time. The first injection (always s.c.) was in the presence of complete Freund’s adjuvant, and the subsequent ones with incomplete Freund’s adjuvant. The antibodies were prepared from blood collected after 7 weeks of immunization. Similar peptides were identified by the same ELISA method, but with a panel of six different monoclonal antibodies prepared as previously reported (H&ion et al., 1995). Lethality tests
A limited number of lethality tests were conducted with mice, crustaceans and crickets. Two albino mice (strains CDl) were injected intraperitoneally with 200 ng of NTX2 per 20 g mouse. None of the animals showed any symptoms of intoxication. Similarly, in our crustacean assay (sweet water shrimp), 30 pg of NTX per g of body weight of crustacean did not affect the animals. However, adult crickets injected with 30 pg of NTX2 showed symptoms of intoxication: paralysis during the first 15 min, followed by a clear hind limb paralysis and uncoordinated movements, but recovered after a couple of hours. Further lethality tests were prevented due to the limited amounts of highly purified NTX2 Amino acid analysis and sequencing
Samples (about 1 nmole each) of the pure native NTX2, and their fragments generated by enzymatic digestion, were analyzed in a Beckman 6300E amino acid analyzer, after acid hydrolysis for 20 hr in 6 N HCI at I1 0°C. An ahquot of pure NTX2 (100 fig) was reduced and alkylated with iodoacetic acid, as described (Martin et af., 1994). An automatic ProSequencer (Millipore model 6400/6600) was used in order to determine the amino acid sequence of 1. native NTX2; 2. its reduced and carboxy-methylated derivative (RC-toxin); and 3. their HPLC-purified fragments produced by trypsin hydrolysis and cyanogen bromide cleavage Two types of polyvinylidene difluoride membranes derivatized with aryl-amine (AM-membranes) and 1,4-phenylene diisothyocyanate (DITC-membranes) were used to attach the peptides for sequencing, as recommended by the manufacturer (Millipore Co., Milford, MA, U.S.A.). Trypsin from Boehringer (Mannheim, Germany) was used at I:50 ratio dilution (trypsin:NTXZ) in the presence of 20 mM Tris-HCI buffer pH 7.8, for 12 hr, at room temperature, while cyanogen bromide (I mg/ml) cleavage of RC-toxin (30 pg) was obtained by incubation in 0.4 ml of 70% formic acid, at room temperature, overnight. All peptides were purified by HPLC prior to sequence determination. Protein content
The concentration of peptides used for binding and electrophysiological on amino acid analysis.
measurements were estimated based
Binding and displacement assays
NTX was radiolabeled with [125]iodineand used for binding and displacement assays, as described (Valdivia membranes (fraction P3) were prepared by the method of Catterall et al. (1979), and used for assessing the capability of both NTX and NTX2 to displace the binding of [‘Z51]NTX (Valdivia et al., 1992).
et al., 1992). Rat brain synaptosomal
Electrophysiological experiments
Bovine aortic endothelial cells were cultured and assayed as described (Vaca et al., 1993). The outside-out configuration of the patch clamp (Hamili et a/., 1981) was used to study single channels obtained from excised patches from single endothelial cells. The recordings and single channel analysis (PO= open time/total time) were performed also, as described (Vaca et al., 1993).
RESULTS AND DISCUSSION
Peptide pur$cation
and amino acid sequence
A novel peptide was purified from the venom of the scorpion C. noxius by essentially the same procedure originally described (Possani et al., 1982), with the addition of two
916
A. R. NIETO et al.
HPLC steps. Briefly, the soluble venom was separated by gel-filtration chromatography in Sephadex G-50 [data identical to Possani et al. (1982)]. Fraction II from this column was applied to a CM-cellulose column providing the separation shown in Fig. l(a). Sub-fraction 1lb, which contains the K+-toxins as indicated by positive ELISA with anti-NTX antibodies, was further selected for additional separation on HPLC as shown
Volume 1.c
( ml 1
b iQS-
a LI OO
/ /
./I 0
/
/ //
50
mm o\”
'12
//
, 20
I
20 4Q,’ Time (/min) / / /
3 L
40
)
Time (min) Fig. 1. Ion-exchange separation of fraction 11. (a) A CM-Cellulose column (0.9 x 30 cm) was loaded with 100 mg of fraction II from Sephadex G-50 separation of C. noxius venom, in the presence of 20 mM ammonium acetate buffer, pH 4.7 and eluted with a linear gradient from 0 to 0.5 M of NaCl in this buffer. The flow rate was 30 ml/hr and tube-fractions containing 2.5 ml were collected. (b) Sub-fraction 11b was further applied to a Cl8 reverse-phase column of a HPLC apparatus and separated with a linear gradient from solution A (0.12% trifluoroacetic acid in water) to 60% solution B (0.10 trifluoacetic acid in acetonitrile). Component 12 was finally applied to a C4 reverse-phase column [inset Fig. l(b)] and eluted with a constant solvent concentration of 40% solution B.
Noxiustoxin 2, a Novel K+ Channel Blocking Peptide
917
in Fig. 1(b). The first HPLC separated about 20 components, the largest peak corresponds to NTX, and component 12 with an elution time of 36.25 min contained NTX2. Since this chromatogram showed some overlapping traces, component 12 was re-applied into a C4 reverse-phase analytical column, using a step gradient of 40% acetonitrile in water, containing 0.1% trifluoroacetic acid [see inset Fig. l(b)]. A couple of small contaminants were eliminated, and the main component was pure toxin NTX2, which was used for chemical and physiological characterization. Final recovery of NTX2 shows that it is 0.13% of the soluble venom. The identification of the NTX-like peptides was performed with polyclononal antibodies. Figure 2(a) shows the positive HPLC fractions recognized in the ELBA test. Peptides eluting at eight distinct retention times gave a positive result (numbers 2, 5, 6, 8, 9, 10, 12, and 15), using polyclonal anti-NTX antibodies. A little < 50% recognition was obtained with component 12, when compared to the control [NTX in Fig. 2(a)]. After
12345
NTX
67
NTX-2
0
’
3 10 11 12 13 14 15 16 NTX
NTX
NTX-2
Fig. 2. ELISA recognition of fractions from HPLC. (a) Percentage of positive recognition of sub-fractions from Fig. l(b), by polyvalent antibodies against NTX, using an ELISA format (Martin et al., 1994). Fractions correspond to the following elution times (in minutes shown in parentheses): 1 (23.90); 2 (24.74); 3 (25.19); 4 (25.26); 5 (28.32); 6 (29.76); 7 (31.04); 8 (31.85); 9 (33.26); 10 (34.62); 11 (35.63); 12 (36.26); 13 (36.77); 14 (37.92); 15 (38.11); 16 (39.90). The last is NTX used as control, and fraction number 12 corresponds to NTX2. (b) Similar results obtained with NTX (control) and NTX2 after fully purified by step-gradient HPLC [inset Fig. l(b)]. (c) Similar experiments using a set of six different monoclonal antibodies against NTX (left side): clones BNTX4, - 12, - 14, - 16, - 18, and - 21 from H&ion er al. (1995) and the same monoclonals (right side) against NTX2 (in that order, from left to right side).
918
A. R. NIETO et al.
1
10
TIINEKCFAT
:------d-m :.T2.>:... :======r== :e l =*=.=*=
20 30 38 SQCWTPCKKA IGSLQSKCMN GKCKCYNG ---me> :- .-.cb-> Tl . . ...>.. ..T4..> :.T3> st===dltc= ------------__---, .=.=.am=,= .=*=*=*=*3 .=.> aa
Fig. 3. Primary structure determination of NTX2. Numbers on top indicate the amino acid position in the sequence (one letter code). Underlined are the sequences obtained with direct automatic Edman degradation (-d-), residues l-16; cyanogen bromide cleaved peptide (.-cb.-), residues 29-37; tryptic fragments (.Tl., eluted at 32.50 min on the HPLC; .T2., at 23.74; .T3., at 26.46; and .T4., at 24.43 min) corresponding to residues from 7 to 18, 1 to 6, 33 to 37, and 18 to 27, respectively; RC-toxin attached to DITC membranes ( = ditc = ), residues l-30; and, finally RC-toxin attached to aryl-amine membranes ( = .am = .), residues l-34. Amino acid determined by amino acid analysis is indicated by aa.
the second HPLC separation, inset Fig. l(b), the pure NTX2 was still positive in the ELISA assay, as shown on Fig. 2(b), while the same homogeneous toxin was not recognized by a panel of six different monoclonal antibodies raised against NTX [Fig. 2(c)]. Figure 3 shows the results of the amino acid sequence of NTX2. Direct sequencing of native toxin permitted the identification of the first 15 residues, while the analysis of RC-toxin attached to aryl-amine- and DITC-membranes provided identification of up to residue number 35. Cyanogen bromide cleavage was important to confirm the C-terminal fragment, from Met29 to Asn37. Several overlapping fragments were obtained by tryptic digestion of RC-toxin, as indicated in Fig. 3. The last residue was surmised based on the results of the amino acid analysis of both the native NTX2 and the C-terminal fragment produced by cyanogen bromide cleavage. Thus, NTX2 is a 38 amino acid residue peptide containing six half-cystines. These results may suggest the presence of three disulfide bridges, as demonstrated for the other known K +-toxins (Possani et al., 1982; Menez et al., 1992; Garcia et al., 1995). Recognition of K+ channels Since NTX2 was 61% identical to NTX, it was interesting to determine the ability of NTX2 to recognize KC-channels. Two preparations were used, as shown in Fig. 4. By means of binding and displacement experiments, using radiolabeled NTX (Valdivia et al., 1992) it was found that both native NTX and NTX2 were effective in competing for similar binding sites (K+ channels) of membranes obtained from rat brain synaptosomes [Fig. 4(a)]. Interestingly, while cold NTX displaced radiolabeled-NTX with a K, close to 100 pM, NTX2 was able to displace 50% binding only at 100 nM. In other words, NTX2 has a thousandfold less affinity than NTX, for the KC-channels of synaptosomes. A second preparation, the outside-out configuration of the patch-clamp method [Fig. 4(b)], was used to study the effect of NTX and NTX2 on single Ca2+-activated K+ channels of small conductance obtained from cultured bovine aortic endothelial cells (Vaca et al., 1993). In this preparation, the 50% probability of opening (P,,) of these channels was decreased by at least two orders of magnitude when comparing NTX and NTX2, the second being less effective than the first one. Structure-function relationship Why would the scorpion venom contain more than one such toxin? And why are their affinities so different? If the known K+ -toxins are compared (Table 1) it is evident that
Noxiustoxin
2, a Novel
0-
K’
Channel
-12-11-10-9
Blocking
-6 -7-6
Log [peptide]
Peptide
919
-5
M
1.0
b
0.0
PO 0.4 0.2 0.6 ~ 0.0
-10-6-6-4-2
0
Log [peptide]
M
Fig. 4. Binding properties and channel recognition of NTX2. (a) Binding and displacement experiments were performed using brain synaptosomal membranes (85 ng per assayed-point) incubated with [‘2’I]Noxiustoxin (50 pM, specific activity 2100 Ci/mmole) and displaced with increasing concentration of cold NTX (0) and NTX2 (0). Mean value of triplicated measurements are plotted. (b) Concentration-response curve for NTX (0) and NTX2 (0) that affected the obtained probability of opening (PJ of single Ca2 + activated K + channels of small conductance from cultured bovine aortic endothelial cells. Mean value of quadruphcated measurements are shown.
there is an important variation at the level of their primary structures, which range from 26 to 61% identity to NTX2. The American species from the genus Centruroides and Tityus have about 50-61% identity, while the scorpions from the Old World of the genus Leiurus, Androctonus and Buthus have only 2644% identity with that of NTX2. However, the most important point is that out of these 40 positions shown in Table 1 ( - 1 position was included in order to increase similarity) only the cysteines and two lysines, at positions 28 and 33 are absolutely conserved. The consensus sequence for all K + -toxins, for which the full sequences are known, presents several conserved residues at the carboxyl-terminal segment. This suggests that the C-terminal segment is important either for function or structural reasons, positioning the N- and C-terminal regions in the right conformation for channel recognition. From the three-dimensional NMR-solution of various K+-channel toxins it is known that the N-terminal and C-terminal region are relatively close, maintained by the three disulfide bridges (Bontems et al., 1992; Meunier et al., 1993; Fernandez et al., 1994; Johnson et al., 1994; Krezel et al., 1995). It is also known, at least for the case of ChTX (Park and Miller, 1992) that the Lys27, equivalent to position 28 of Table 1, is a ‘crucial residue’ for the interaction with the Shaker B K + -channels (Stampe
A. R. NIETO et al.
920
Table 1. Comparative amino acid sequence of K+ channel toxins Toxin 1 NTX2 NTX MgTX CllTX 1 TsKa KlTX AgTXl Lq2 AgTX2 AgTX3 IbTX ChTX LeTX I PO5 Consensus
Identity (%)
Amino acid seauence -TIINEKCFAT -TIINVKCTSP -TIINVKCTSP -1TINVKCTSP -VFINAKCRGS GVEINVKCSGS GVPINVKCTGS pEFTQESCTAS GVPINVSCTGS GVPINVPCTGS pEFTDVDCSVS pEFTNVSCTTS -AF----C-NL TV--NL
--
10
20
30
SQCWTPCKKA KQCSKPCKEL KQCLPPCKAQ QQCLRPCKDR PECLPKCKEA PQCLKPCKDA PQCLKPCKDA NQCWSICKRL PQCIKPCKDA PQCIKPCKDA KECWSVCKDL KECWSVCQRL RMCQLSCRSL RRCQLSCRSL -Cd-
IGSLQS-KCM YGSSAGAKCM FGQSAGAKCM FGQHAGGKCI IGKAAG-KCM GMRF+KCM GMRF-G-KC1 HNTNRG-KCM GMRF-G-KCM GMRF-G-KCM FGVDRG-KCM HNTSRG-KCM -GL-LG-KC1 -GL-LG-KC1 KC-
NGKCKCYNG NGKCKCYNN NGKCKCYPH NGKCKCYPNGKCKCYPNRKCHCTPK NGKCHCTPK NKKCRCYSNRKCHCTPK NRKCHCTPK GKKCRCYQNKKCRCYSGDKCECVKH GVKCECVKH -KC-C-
39 100 61 56 51 50 44 44 42 41 41 37 34 26 26
NTXZ, from this work; CllTX 1, toxin 1 from Martin et al. (1994); NTX, noxiustoxin from Possani et al. (1982); MgTX (margatoxin), ChTX (charybdotoxin), IbTX (iberiotoxin), Lq2 (L. quinquestriatus toxin 2) and AgTXl to AgTX3 from L. quinquestriatus var. hebraeus [reviewed by Garcia et al. (1995)]; LeTX I, leirutoxin 1 (Auguste et al., 1990; Chicchi et al., 1988). KlTX, kaliotoxin (Crest et al., 1992); TsKa from the venom of T. serrulatus (Rogowski et al., 1992), and PO5 from the venom of A. mauretanicus mauretanicus (Sabatier et al., 1993). Consensus stands for positions in which all the amino acids are conserved in all sequences. Gaps (-) were introduced to enhance similarities.
et al., 1994). However, for NTX the N-terminal amino acid sequence is very important. Actually, a nonapeptide corresponding to this section is toxic to mice (Gurrola et al., 1989) and recognizes some sub-types of K+ channels (Vaca et al., 1993). NTX2 presents two highly similar amino acid sequences to NTX, one being the N-terminal (first eight residues, except for position 5) and the other at the C-terminal segment (positions 28-38). The presence of a negatively charged amino acid (Glu) in position 5, substituting for the hydrophobic valine in NTX, probably renders this segment unfavorable for binding to the K + channels, as shown in Fig. 4. Thus, functional differences shown to exist between NTX and NTX2 might reside preferentially, either at the important changed glutamic acid of position 5, or at some other highly variable region of the internal segment. However, this does not rule out the possibility that both the N-terminal and the internal segments (corresponding to positions 9-27), are important for the functional differences observed. Furthermore, worth mentioning is the lack of glycine at position 27 which was claimed to be essential for proper folding of ChTX (Bontems et al., 1992; MCnez et al., 1992). This residue is present in all the other K+-toxins, but is missing in NTX2. In conclusion, it is quite possible that all the variations we observed in terms of primary structure, despite the fact of a common structural motif at the three-dimensional level, (Menez et al., 1992) are necessary modifications evolved by scorpions to cope with the co-evolving types and sub-types of K+ -channels. This might answer the question why scorpions have so many different toxins in the same venom. Additionally, we should consider the possibility of species specificity discussed earlier. Some of these peptides are designed to recognize different tissues from different taxonomic groups, like the case of NTX2, which is toxic only to insects.
Noxiustoxin 2, a Novel K+ Channei Blocking Peptide
921
Acknowledgements-The technical assistance of M.Sc. Fernando Z. Zamudio, Timoteo C. Olamendi-Portugal, and Sandra Contreras is greatly acknowledged. Supported in part by grants from Howard Hughes Medical Institute (No. 75191-527104), from National University of Mexico (DGAPA No. IN-205893 and IN-206994), National Council of Research and Technology of Mexico (CONACyT No.4734N), and European Commission (CII*-CT 94-0045). A. Nieto received a scholarship from CONACyT.
REFERENCES Auguste, P., Hugues, M., Grave, B., Gesquitre, J. C., Maes, P., Tartar, A., Romey, G., Schweitz, H. and Lazdunski, M. (1990) Leiurotoxin I (Scyllatoxin), a peptide ligand for Ca’+ -activated K+ channels. /. hiol. Chem. 265, 47534759.
Bontems, F., Gilquin, B., Roumestand, C., Mtnez, A. and Toma, F. (1992) Analysis of side-chain organization on a refined model of charybdotoxin: structural and functional implications. Biochemistry 31, 77567764. Carbone, E., Wanke, E., Prestipino, G., Possani, L. D. and Maelicke, A. (1982) Selective blockage of voltage-dependent K’ channels by a novel scorpion toxin. Nature (London) 2%, 90-91. Catterall, W. A., Morrow, C. S. and Harfshorne, R. P. (1979) Neurotoxin binding to receptor sites associated with voltage-sensitive sodium channels in intact, lysed and detergent solubilized brain membranes. J. hiol. Chem. 254, 11379-11387. Chicchi, G. G., Gimenez-Gallego, G., Ber, E., Garcia, M. L., Winquist, R. and Cascieri, M. A. (1988) Purification and characterization of a unique, potent inhibitor of apamin binding from Leiurus quinquestriatus hebraeus venom. J. biol. Chem. 263, 10192-10197. Crest, M., Jacquet, G., Gola, M., Zerrouk, H., Benslimane, A., Rochat, H., Manuelle, P. and Martin-Eauclaire, M. F. (1992) Kaliotoxin, a novel peptidyl inhibitor of neuronal BK-type Ca*+-activated K+ channel characterized from Androctonus mauretanicus mauretanicus venom. J. biol. Chem. 267, 164&1647. Fernandez, I., Romi, R., Szendeffy, S., Martin-Eauclaire, M. F., Rochat, H., Van Rietschoten, J., Pons, M. and Giralt, E. (1994) Kaliotoxin (l-37) shows structural differences with related potassium cahnnel blockers. Biochemistry
33, 1425614263.
Garcia, M. L., Knaus, H. G., Munujos, P., Slaughter, R. and Kaczorowski, G. J. (1995) Charybdotoxin and its effects on potassium channels. Am. J. Physiol. (Cell Physiol) 269, Cl-ClO. Gurrola, G. B., Molinar-Rode, R., Sitges, M., Bayon, A. and Possani, L. D. (1989) Synthetic peptides corresponding to the sequence of Noxiustoxin indicate that the active site of this K + channel blocker is located on its amino-terminal portion. J. Neur. Transmission 77, 1l-20. Gurrola, G. B. and Possani, L. D. (1995) Structural and functional features of Noxiustoxin: a K” channel blocker. Biochem. Mol. Biol. Int. 37, 527-535. Hamill, 0. P., Marty, A., Neher, E., Sakamann, B. and Sigworth, F. J. (1981) Improved patch clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pfiigers Arch. 391, 855100. Herion, P., Gurrola-Briones, G., Sanchez, M. R., Saavedra, R. and Possani, L. D. (1995) Monoclonal antibodies against Noxiustoxin. Hybridoma 14, 247-251. Hille, B. (1992) Mechanisms of block. In: Ionic Channels of Excitable Membranes, pp. 39&422. (Hille, B., Ed.). Sunderland.: Sinauer Assoc. Inc. Johnson, B. A., Stevens, S. P. and Williamson, J. M. (1994) Determination of the three-dimensional structure of margatoxin by ‘H, “C, “N triple resonance nuclear magnetic resonance spectroscopy. Biochemistry 33, 15061-15070.
Knaus, H. G., Koch, R. O., Eberhart, A., Kaczorowski, G. J., Garcia, M. L. and Slaughter, R. S. (1995) [Y]Margatoxin, an extraordinarily high affinity ligand for voltage-gated potassium channels in mammalian brain. Biochemistry 34, 13627-13634. Krezel, A. M., Kasibhatla, C., Hidalgo, P., MacKinnon, R. and Wagner, G. (1995) Solution structure of the potassium channel inhibitor agitoxin 2: caliper for probing channel geometry. Protein Sri. 4, 1478-1489. Martin, B. M., Ramirez, A. N., Gurrola, G. B., Nobile, M., Prestipino, G. and Possani, L. D. (1994) Novel K + -channel blocking toxins from the venom of the scorpion Centruroides limpidus ltmpidus Karsch. Biochem. J. (Great Britain)
384, 5166.
Mtnez, A., Bontems, F., Roumestand, C., Gilquin, B. and Toma, F. (1992) Structural basis for functional diversity of animal toxins. Proc. R. Sot. Edinburgh 99B, 83-103. Meunier, S., Bernassau, J. M., Sabatier, J. M., Martin-Eauclaire, M. F., Van Rietschoten, J., Cambillau, C. and Darbon, H. (1993) Solution structure of POrNH2, a scorpion toxin analog with high affinity for the apamin-sensitive potassium channel. Biochemistry 32, 11969-l 1976. Park, C. S. and Miller, C. (1992) Mapping function to structure in a channel-blocking peptide: electrostatic mutants of charybdotoxin. Biochemistry 31, 7749-7755. Possani, L. D. (1984) Structure of Scorpion Toxins. In: Handbook of Natural Toxins, Vol. 2, pp. 513-550 (Tu, A. T. Ed.). New York: Marcel Dekker. Possani, L. D., Martin, B. M. and Svendsen, I. B. (1982) The primary structure of Noxiustoxin: a K’ Channel blocking peptide from the venom of the scorpion Centruroides noxius Hoffmann. Curlsberg Res. Commun. 47, 285-289.
922
A. R. NIETO et al
Rogowski, R. S., Collins, J. H. and Blaustein, M. P. (1992) Purification and structural characterization of a potent potassium channel blocker from scorpion Tityus serrulatus venom. Biophys. J. 61, A250 (abstract). Sabatier, J. M., Zerrouk, H., Darbon, H., Mabrouk, K., Benslimane, A., Rochat, H., Martin-Eauclaire, M. F. and Van Rietschoten, J. (1993) POS, a new Leiurotoxin I-like scorpion toxin: synthesis and structure-activity relationships of the alpha-amidated analog, a ligand of Ca’+ -activated K + channels with increased affinity. Biochemistry
32, 2163-2170.
Stampe, P., Kolmakova-Partensky, L. and Miller, C. (1994) Intimations of K+ channel structure from a complete functional map of the molecular surface of charybdotoxin. Biochemistry 33, 443450. Vaca, L., Gurrola, G. B., Possani, L. D. and Kunze, D. (1993) Blockade of a Kc. channel with synthetic peptides from Noxiustoxin: a K+ channel blocker. J. membr. Biol. 134, 123-129. Valdivia, H. H., Smith, J. S., Martin, B. M., Coronado, R. and Possani, L. D. (1988) Charybdotoxin and noxiustoxin, two homologous peptide inhibitors of the K + (Ca”) channel. FEBS Lett. 226, 280-284. Valdivia, H. H., Martin, B. M., Escobar, L. and Possani, L. D. (1992) Noxiustoxin and leiurutoxin III, two homologous peptide toxins with binding properties to synaptosomal membrane K + channels. Biochem. Int. 27, 953-962.
Vazquez, J., Feigenbaumm, P., King, V. F., Kaczorowski, G. J. and Garcia, M. L. (1990) Characterization of high affinity binding sites for charybdotoxin in synaptic plasma membranes from rat brain. J. biol. Chem. 265, 15564-15571.