Possani2 1 Institute of Chemistry, National Autonomous University of Mexico, Mexico City, Mexico;2Department of Molecular Medicine and Bioprocesses, Institute of Biotechnology, National
Trang 1NMR solution structure of Cn12, a novel peptide from the Mexican
but with a-like physiological activity
Federico del Rı´o-Portilla1, Elizabeth Herna´ndez-Marı´n1, Genaro Pimienta2*, Fredy V Coronas2,
Fernando Z Zamudio2, Ricardo C Rodrı´guez de la Vega2, Enzo Wanke2,3and Lourival D Possani2
1
Institute of Chemistry, National Autonomous University of Mexico, Mexico City, Mexico;2Department of Molecular Medicine and Bioprocesses, Institute of Biotechnology, National Autonomous University of Mexico;3Dipartimento di Biotecnologie e Bioscienze, Universita´ di Milano-Bicocca, Milan, Italy
Cn12 isolated from the venom of the scorpion Centruroides
noxiushas 67 amino-acid residues, closely packed with four
disulfide bridges.Its primary structure and disulfide bridges
were determined.Cn12 is not lethal to mammals and
arthropods in vivo at doses up to 100 lg per animal Its 3D
structure was determined by proton NMR using 850
dis-tance constraints, 36 / angles derived from 36 coupling
constants obtained by two different methods, and 22
hydrogen bonds.The overall structure has a two and half
turn a-helix (residues 24–32), three strands of antiparallel
b-sheet (residues 2–4, 37–40 and 45–48), and a type II turn
(residues 41–44).The amino-acid sequence of Cn12
resem-bles the b scorpion toxin class, although patch-clamp
experiments showed the induction of supplementary slow inactivation of Na+channels in F-11 cells (mouse neuro-blastoma N18TG-2· rat DRG2), which means that it behaves more like an a scorpion toxin.This behaviour prompted us to analyse Na+channel binding sites using information from 112 Na+channel gene clones available in the literature, focusing on the extracytoplasmic loops of the S5–S6 transmembrane segments of domain I and the S3–S4 segments of domain IV, sites considered to be responsible for binding a scorpion toxins
Keywords: Centruroides noxius; NMR structure; patch-clamp; scorpion toxin; sodium channel
Scorpion toxins are relatively short peptides with a variable
length of amino acids, showing characteristic 3D folding
comprised of an a-helix and three segments of antiparallel
b-sheet structure, stabilized by several disulfide bridges
[1–5].Their known physiological role is to block or modify
ion-channel function, causing impairment of cellular
com-munication, which leads to the depolarization of excitable
membranes and might cause death of animals stung by
scorpions [6,7]
There are several reasons why the molecular basis of
toxin specificity and molecular mechanism of action
continue to be of scientific interest: (a) to study ion channels,
the target molecules of most known scorpion toxins, in
order to understand their molecular structure and function,
thus learning more about cellular excitability; (b) to
understand the toxic effects of scorpion venoms, a
pre-requisite for the development of more effective and safer antidotes and/or vaccines; (c) to find toxins specific for invertebrate organisms with a view to developing biode-gradable drugs for pest control; (d) to discover other possible unknown target molecules for which peptides were evolved in the venom of scorpions.The last of these is not trivial, as the huge variability of these peptides, estimated to
be of the order of 100 000 in scorpion venom alone, and of which only about 0.2% have been identified, leaves a wide open field for research [4,5,8].Several recent articles and reviews have reported on the structural and functional aspects of these peptides [3,8–18].Most dealt with scorpion toxins as ion-channels blockers or modifiers of their function.However, the structural variability of peptides and the different types of receptor they recognize is steadily increasing [19], exemplified by the following novel discov-eries: ERG-channel-specific toxins [20], analgesic peptides [21,22], modulators of immune response [23–25], antibiotics [26–28], antimalaria agents [29], and others
For the ion-channel-specific peptides, two distinct groups
of toxins have been identified based on the length of the peptide chain: short-chain peptides (23–41-amino-acids long) which recognize and bind to various types and subtypes of K+channels [13,19], ryanodine-sensitive Ca2+ channels [11], and Cl– channels [30]; long-chain peptides (59–76 amino acids), which are specific for Na+channels [8,18] and T-type Ca2+channels [31,32]
The purpose of this paper is to describe for the first time a Na+-channel-specific scorpion toxin (Na-ScTx), isolated from the New World scorpion Centruroides
Correspondence to L.D.Possani, Department of Molecular Medicine
and Bioprocesses, Institute of Biotechnology, National Autonomous
University of Mexico, Avenida Universidad 2001, Apartado Postal
510-3, Cuernavaca 62210, Mexico.Fax: + 52 777 3172388,
Tel.: + 52 777 3171209, E-mail: possani@ibt.unam.mx
Abbreviation: Na-ScTx, scorpion toxin specific for Na + channel; TTX,
tetrodotoxin.
Note: Laboratories 1 and 2 contributed equally to this work.
*Present address: EMBL, Meyerhofstrasse 1, Heidelberg D-69117,
Germany.
(Received 4 February 2004, revised 30 March 2004,
accepted 22 April 2004)
Trang 2charge distribution on the surface of the toxin for its
activity is emphasized.Furthermore, considerations
rela-ted to the amino-acid sequences deduced from cloned
Na+-channel genes are discussed in terms of what is
known about the interacting surfaces of Na-ScTxs and
Na+channels
Materials and methods
Venom source, purification procedures and lethality tests
Venom from scorpions collected in Nayarit State (Mexico)
was obtained by electrical stimulation and separated by
Sephadex G-50 gel filtration, followed by ion-exchange
chromatography on CM-cellulose columns, as previously
described [33] For this work, fraction II-4 (v.g., fraction II
from Sephadex, and subfraction 4 from CM-cellulose) was
further separated by HPLC using previously published
conditions [34].Lethality tests were conducted with mice,
crickets and crayfish, using the same conditions as reported
[34]
Determination of amino-acid sequence and MS analysis
The full amino-acid sequence of the toxin was obtained by
direct Edman degradation, using an automatic Beckman
sequencer (LF 3000 Protein Sequencer) and samples of
native and reduced toxin.Additional information was
generated by sequencing subpeptides obtained by HPLC
separation of toxin treated with endopeptidases lysine-C
(Boehringer, Mannheim, Germany) and Staphylococcus
aureus V8 (Boehringer, Mannheim, Germany), as
previ-ously described [8,33,34].The last amino-acid residue was
confirmed by MS analysis, performed in a LCQDUO
ion-trap spectrometer from Finnigan (San Jose, CA, USA)
Determination of disulfide bridges
A sample containing 100 lg toxin was digested with
trypsin (Promega, Madison, WI, USA) in slightly acidic
conditions, followed immediately by endopeptidase V8
digestion, as previously described [8], and separated by
HPLC on a C18 reverse-phase column with a linear
gradient from solution A [0.12% (v/v) trifluoroacetic acid
in water] to 60% solution B [0.10% (v/v) trifluoroacetic
acid in acetonitrile], run for 60 min.Several components
were fractionated (data not shown).The component
eluted at 27.78 min was identified as the disulfide bridge
between Cys11 and Cys65.Because of incomplete
digestion, the peptide eluted at 28.92 min was further
treated with chymotrypsin (Boehringer) and subsequently
separated by HPLC.The component eluted at 26.38 min
(data not shown) corresponded to the disulfide bridge
between Cys15 and Cys40.The two other disulfide
bridges were determined by NMR analysis as discussed
below.For the assignment of the first two disulfide
bridges, the peptides hydrolyzed by enzymatic digestion
were sequenced using the Beckman LF3000 Protein
Sequencer described above
N18TG-2· rat DRG) [35] were routinely cultured in Dulbecco’s modified Eagle’s medium, containing 4.5 gÆL)1 glucose and 10% fetal calf serum.The cells were incubated
at 37C in a humidified atmosphere with 5% CO2 Solutions and drugs The standard extracellular solution contained (mM): NaCl 130, KCl 5, CaCl2 2, MgCl2 2, Hepes/NaOH 10,D-glucose 5, at pH 7.40 In the high-K+ external solution ([K+]o¼ 40 mM), NaCl was replaced by
an equimolar amount of KCl.The standard pipette solution
at [Ca2+]i¼ 10–7
M(pCa 7) contained (mM): K+-aspartate
130, NaCl 10, MgCl22, CaCl21.3, EGTA/KOH 10, Hepes/ KOH 10, ATP (Mg2+salt) 1, pH 7.30
Patch-clamp recordings and data analysis The currents were recorded by means of the patch-clamp amplifier MultiClamp 700A (Axon Instruments, Foster City, CA, USA) at room temperature as previously described [36] (pipette resistance 0.8–1.2 MW); cell capacitance and series resistance errors were carefully compensated for (85–90%) before each voltage clamp protocol run.The extracellular solutions were delivered through a nine-hole (0.6-mm) remote-controlled linear positioner, with an average response time of 2–3 s, placed near the cell under study The Na+-current inactivation curves were obtained by plotting the normalized peak current against Vm.Final traces were corrected with traces obtained in the presence of
20 nMtetrodotoxin (TTX).The activation was derived as the normalized sodium conductance relationship (ENa¼ )65).To determine the amplitude of the toxin-induced slow inactivation, we subtracted the control traces from the traces recorded in the presence of the toxin.The amplitude of these toxin-induced currents was analyzed by plotting the value
3 ms after the onset of the depolarizing pulse.The decaying inactivating portion of the control traces and the currents in the presence of toxin were fitted to one or two exponential decaying functions, respectively, to obtain the inactiva-tion time constants.pClamp 8 (Axon Instruments) and Origin 4.1 (Microcal Inc, Studio City, CA, USA) software were routinely used during data acquisition and analysis Preparation of the NMR sample
A sample of purified Cn12 containing 6.2 mg peptide was dissolved in 0.8 mL H2O/D2O (9 : 1, v/v) to a final concentration of 0.9 mM.The pH measured for this solution was 3.1 After the experiments in H2O, the peptide was lyophilized and redissolved in D2O to perform additional experiments
1 H-NMR spectroscopy The experiments were performed in a Varian Unity Plus 500 spectrometer.Data were collected at 300 K.Mixing times were 35 ms for TOCSY and 80 and 100 ms for NOESY in
H2O and in D2O.Data were processed withNMRPIPE[37] to obtain 4K· 4K spectra.Spectra were analysed using the XEASY program [38].The values of the JHN-Ha were estimated from TOCSY spectra with the modified
Trang 3Jdoubling in the frequency domain [39,40], when possible,
or with the strategy proposed by Wishart et al.[41]
Experimental constraints and structure calculations
Most of the distance constraints were obtained from
NOESY spectra in H2O; additional constraints were from
NOESY spectra in D2O.NOE intensities were evaluated
from the volume of the cross-peak and calibrated internally
using theCALIBAprogram [42] to generate a set of upper
limit distances.Most NOE data were obtained from
resolved signals.NOE signals for Hb–Hbbetween Cys11
and Cys65 and Cys15 and Cys40 were assigned; however,
no NOE data were observed for the other Cys pairs
A total of 850 distance constraints were used from which
121 are sequential, 30 medium range (1 < |i–j| < 4) and
92 long range (|i–j| > 4); the remainder were intraresidues
A total of 36 / angle constraints were used based on the
JHN-Havalues.Using the modified J doubling method, it
was possible to evaluate 23 JHN-Havalues from a trace of
TOCSY spectra.It was only possible to measure 13
additional coupling constants from TOCSY spectra as
proposed by Wishart.It was found that the modified
Jdoubling method gives smaller values than the Wishart
method [41], and in the case of the a-helix (residues
24–32), where a value of 3.9 is expected [43], values are
closer depending on the structure obtained.Proton–
deuterium exchange of the amide groups was measured
on a sample lyophilized from H2O and redissolved in pure
D2O as described [43], in order to determine 22 hydrogen
bond constraints.Four disulfide bridge constraints were
added to the calculations.The dynamic annealing
struc-ture calculations were performed with the CNS software
suite [44]
Analysis of Na+-channel sequences
By searching data banks containing sequences of Na+
channels from human, fruit fly and squid, several
representative sequences were chosen (five for humans,
three for Drosophila, and one for squid).With these
sequences, more than 200 BLASTentries were found that
matched isoforms of Na+-channel sequences.An
align-ment was performed using the program CLUSTAL X [45],
from which 50 nonredundant sequences were selected
Among these are sequences from several species including
mammals, insects and other invertebrates (available from
the authors on request).The proposed binding sites for
a Na-ScTxs, including regions corresponding to the S5–S6
segment from domain I and S3–S4 from domain IV of
the Na+ channels, were re-aligned independently using
CLUSTAL X, and some were chosen for our figure.Among
these are the ones best represented and cited in the
literature on scorpion toxins
Results
Purification and sequencing
Figure 1A shows the results of separating 1 mg fraction II-4
(for details see Materials and methods) from the venom of
the scorpion C noxius.For this report, about 20 HPLC
runs were performed as described, in order to obtain enough peptide for this work.The component indicated by the star
in Fig.1A was recovered and chromatographed again (Fig.1A, inset) to give the pure peptide, called Cn12.The name comes from the abbreviation of the scorpion species
C noxiusfollowed by the number 12, which corresponds to the 12th pure peptide fully characterized from this venom, specific for Na+ channels.This peptide corresponds to
1.3% of the total venom protein concentration
For determination of the primary structure, it was necessary to obtain overlapping sequences of several peptides (Fig.1B) The 36 most N-terminal amino-acid residues were identified by directly sequencing the native peptide, and the identities confirmed by sequencing a sample of reduced and alkylated cysteine residues (vinyl derivatives) of Cn12.The overlapping sequence spanning Ala33 to Gly63 was determined with a peptide obtained by Lysine-C digestion, as described in Materials and methods, and the final segment overlapping from Gly53 to Arg66 was obtained by sequencing a peptide digested with endo-peptidase V8.The last residue, Ser in position 67, was determined by MS.The molecular mass found by MS ionization analysis was 7139.5 Da, and the theoretically expected value was 7139.7 Da, thus confirming the full sequence
Disulfide bridges
As mentioned in Materials and methods, a peptide obtained
by enzymatic digestion of Cn12, and separated by HPLC (elution time 27.78 min) provided a heterodimeric pro-duct when directly sequenced by Edman degradation The sequences obtained were DGYPLASNGC and GTVLWGDSGTXPCR, indicating unequivocally the posi-tion of one of the disulfide bridges, which in this case was between Cys11 and Cys65, based on the known primary structure of Cn12 (Fig.1B).Another core peptide from the initial digestion (time 28.92 min) gave several amino-acid sequences, making it impossible to assign any other disulfide bridges.It was therefore further digested with chymotryp-sin, which produced subpeptides, and that eluted at 26.35 min (data not shown) gave the sequences FGC and GYC, corresponding to the disulfide bridge Cys15 and Cys40.The peptide eluted at 17.48 min gave four sequences corresponding to segments of the primary structure con-taining Cys25, Cys27, Cys45 and Cys47.As we knew the primary structure, we knew that the remaining disulfide bridges were between Cys25 and Cys45 and Cys29 and Cys47, or conversely between Cys25 and Cys47 and Cys29 and Cys45.The sequences of the fragments from the chymotryptic hydrolysis showed that the other putative disulfides, i.e Cys25–Cys29 and Cys45–Cys47, were not possible because the sequence would not fit the one determined experimentally
Modelling the 3D structure, based on the NMR data as discussed below, showed that the only disulfide configur-ation that would fit the results obtained is Cys25–Cys45 and Cys29–Cys47.Furthermore, these disulfide pairs are at exactly the same positions as those described for other Na-ScTxs isolated from several other scorpion species [8].Thus, the disulfide bridges of Cn12 are assumed to be: Cys11– Cys65, Cys15–Cys40, Cys25–Cys45 and Cys29–Cys47
Trang 4Sequence comparison of Cn12 with other Na-ScTxs
A databank search with the sequence of Cn12 as the query
retrieved only scorpion venom-derived peptides classified as
Na-ScTxs.The highest identities were found with scorpion
toxins belonging to the b group (identity > 30%), including
several toxins isolated from New World scorpions.It also
shows similarities to some depressant and excitatory toxins
from Old World scorpions, as well as to the recently
characterized Birtoxin, a long-chain peptide with only three
disulfide bridges from the scorpion Parabuthus transvaalicus
[46].The similarities to Na-ScTxs of the a group are
considerably lower (identity < 25%).On the basis of these
results, Cn12 should be classified as a member of the b
group of Na-ScTxs.To allow proper discussion in the terms
of structure–activity relationship, with regard to recognition
and affinity for Na+channels, Cn12 is aligned with other
Na-ScTxs in Fig.2.For this comparison, only toxins for
which the 3D structures are known were included.The 3D
structures of three pharmacologically different classes of
Na-ScTxs are known: a Na-ScTxs (AaHII, BmKM1,
BmKM2, BmKM4, BmKM8, BmK-aTx16, BmK-aIT,
Bs-mktx, CsE-V, LqhII, LqqIII, Lqh-aIT), b Na-ScTxs
(Cn2, CsE1, CsEv1, CsEv2, CsEv3, CsEv5, Ts1/Tsc), and
an insect-specific excitatory Na-ScTx Bjxtr-IT.As expected,
the phylogenetic tree rooted with Bjxtr-IT places Cn12
closer to toxins described as b ScTxs than to the a
Na-ScTxs.A recent proposal by Froy & Gurevitz [47] is that toxins CsEv1 and CsEv3 belong to the group of so-called a¢ Na-ScTxs.However, the lack of pharmacological data prevents proper classification of this last group of toxins [7], and thus they are here referred to as b Na-ScTx, on the basis
of their primary structure
Bioassays and electrophysiological effect of Cn12 Bioassays conducted with pure Cn12 in mice, crickets and sweet-water shrimps, using concentrations up to 100 lg per animal produced inconclusive results.Apparently, at this concentration, Cn12 is not toxic to any of the animals tested.The immediate questions posed were: why is this component, which is present at relatively high concentration
in the venom (Fig.1A), not toxic in vivo? What is the biological function of this novel peptide? To answer these questions, it was decided to verify the effect of Cn12 in vitro using patch-clamp experiments
The TTX-sensitive Na+current present in the tumour cell line F-11 [35] was used to test the properties of the toxin Cn12.Figure 3A shows representative recordings of inward currents elicited according to the protocol shown below The effects of 2.8 lMCn12 are illustrated in Fig.3B.The peak currents are higher in the presence of toxin but the most interesting effect was a net increase in the inactivation time constant.I vs.V plots obtained in the same cell in
Fig 1 Final purification and amino-acid
sequence determination of Cn12 (A) A sample
of fraction II-4 from a CM-cellulose
ion-exchange column [33] containing 1 mg protein
was applied to an analytical C 18 reverse-phase
column and eluted with a linear gradient of
solution A [0.12% (v/v) trifluoroacetic acid in
water] to 60% solution B [0.10% (v/v)
tri-fluoroacetic acid in acetonitrile] run for
60 min.The component labelled with asterisk
corresponds to 33% of fraction II-4.It was
further chromatographed in the same system,
but eluted with a gradient from 10% to 40%
solution B over 40 min.Only the material
eluted under the main area of the peak,
indi-cated by the horizontal line, was collected.It
corresponded to pure toxin, whereas small
contaminants (5%) were eliminated on the
ascending and descending sections of the
chromatogram.(B) Direct Edman
degrada-tion of native peptide and reduced and
alkylated samples provided unequivocal
ami-no-acid sequence identification from Arg1 to
Asp36.Further sequencing peptides obtained
from enzymatic hydrolysis with endopeptidase
Lys-C and Staphylococcus aureus protease V8
allowed us to obtain the overlapping segments
from A33 to G63 and from Gly53 to R66,
respectively, as indicated.The last residue,
S67, was identified by MS.
Trang 5the control and during the application of 0.7 and 2.8 lM
Cn12 are shown in Fig 3C
In the voltage range)20 mV to +20 mV, the decaying
inactivation time course was fitted with one single
expo-nential time constant in the control, but, in the presence of
the toxin, we were forced to add another slower time
constant while maintaining unaltered the fast time constant
used in the control.The ratio of the fast to slow amplitudes
was 0.1 ± 0.02 (n ¼ 4).The complete results of four
experiments are shown in Fig.3D where it can be seen that
the control data (open squares) and the Cn12-induced slow
inactivation component data (open circles) differ by about
one order of magnitude.Normalized voltage-dependent
activation is illustrated in Fig.3E both in the control (open
squares) and during the action of Cn12 (open triangles);
the data show no significant difference
Classical double-pulse inactivation protocols were used to
investigate the voltage-dependent steady-state inactivation
process.In this case also (Fig.3E), data did not differ in the
control (open squares) and in the presence of 2.8 lMCn12
(open triangles).As the toxin induced the development of a
novel slowly inactivating component (Fig.3D), we
subtrac-ted the control traces from the toxin traces and plotsubtrac-ted the
steady state inactivation of this toxin-induced component
(closed squares).This resulted in a right shift by about
12 mV
Overall, these data suggest that Cn12 behaves like a weak
a Na-ScTx because it induces supplementary slow
inactiva-tion of the Na+channel.In other words, Cn12 interferes
with cellular communication at the level of the Na+
channels
NMR solution structure of Cn12
Figures 4, 5 and 6 summarize the most important data
obtained from the NMR analysis of pure Cn12.It was
possible to analyze the NMR data because of well-dispersed
signals obtained at 11.75 T Figure 4A shows the NOE
diagram of sequential and medium range data, the chemical shift index, and coupling constants used for the structure calculation.Figure 4B shows a ribbon diagram of Cn12, in which the most obvious 3D elements are shown.This was possible because of the use of 850 distance constraints, 36 / angles derived from 23 coupling constants measured from TOCSY experiments using the modified J-doubling method and 13 additional coupling constants using the Wishart method [41], plus 22 hydrogen bonds added after the first calculations, and the four disulfide constrains (see Materials and methods).Over 250 structures were calculated, from which 19 with the smallest total energy and no NOE violations greater than 0.2 A˚ and no angle constraints violations greater than 5 were used to draw Fig.4C Table 1 shows the rmsd values from different regions in the peptide and energy mean values of calculated structures.All three prolines were determined in trans position because of the presence of NOE data between the previous Ha with the proline HDXprotons [Ha(i–1) Hc(i)].For each proline, at least one NOE was assigned.Concerning the disulfide bridges mentioned above, if a distinct disulfide pairing constraint is used to calculate the plausible 3D structures, the model obtained does not fit the NMR results well, because several NOE violations are produced and structures with very high total energy are obtained.These data led to the conclusion that the two disulfide bridges not yet directly determined are indeed between Cys25 and Cys45 and Cys29 and Cys47, as mentioned above.Additional details of NMR experimental results of Cn12 can be found in the databanks (PDB entry 1PE4 and BRMB number 5913)
Discussion
Purification, sequence and function of known Na-ScTxs The purification procedure and sequence determination of Cn12 are clearly described in the Results section and require
no further discussion.However, as this paper reports the
Fig 2 Multiple sequence alignment of Na-ScTxs Multiple alignment of amino-acid sequences of all Na-ScTxs for which the 3D structure is known was conducted using CLUSTAL X [45].Amino-acid sequences are followed by the abbreviated name and the corresponding PDB code.The third column displays the identity scores with respect to Cn12.In the right part of the figure a simplified phylogenetic tree is included, for which the amino-acid sequence of Bjxtr-IT (a highly divergent insect-specific excitatory toxin) was used as the root.The tree accurately differentiates between
a and b Na-ScTxs, and the corresponding branches are labelled accordingly.Amino acids shown directly to be important in pharmacological activity by site-directed mutagenesis are in bold, as reported in: Lqh-aIT [18], BmK M1 [55,56] and Bjxtr-IT [63].In the upper and lower parts of the figure, the secondary-structure elements (h, helix; b, sheet) and fully (asterisks) or partially (dots) conserved residues are indicated, respectively.
Trang 6structure and function of the first a-toxin specific for Na+
channels, isolated from C noxius, it is important to review
briefly the field, considering the data and current ideas on
the interacting surfaces of scorpion toxins and Na+
channels
The best studied scorpion toxins with respect to
inter-acting surfaces are the K+ channel-specific toxins
[9,13,15,16,19].Less information is available for those
specific for Na+ channels.The Na-ScTxs were initially
classified as a and b toxins [48,49].The a Na-ScTxs bind to
site 3 of the receptor from vertebrates in a
voltage-dependent manner, whereas the b Na-ScTxs bind to site 4,
producing a shift to more negative potential and the binding
is independent of the membrane potential [18,50,51]
On the basis of electrophysiological recordings and
binding and displacement experiments performed with
different animal models (mammals, insects and crusta-ceans), several subgroups of a, a-like, and b Na-ScTxs have been proposed [7,8,47].Still more recently, another novel toxin, Cn11 from C noxius was described, the first example
of a Na+channel blocker peptide in crustacean prepara-tions, in contrast with all other known Na-ScTxs that are modifiers of channel function [34].However, in reality a general classification of the different types and subtypes of Na-ScTxs is not available, which is due in part to the lack of knowledge about their structural and functional relation-ships.In addition, none of the different types and subtypes
of Na+channel have had their 3D structure determined, unlike the K+channels [52]
Previous studies conducted with both a and b Na-ScTxs showed that small differences such as single amino-acid modifications or deletion of short stretches of sequence were
Fig 3 Biophysical modifications produced by Cn12 on the TTX-sensitive sodium currents of the neuroblastoma cell line F-11 (A-B) Na+current elicited from a holding potential of )100 mV to the test potentials from )30 to +20 according to the protocol shown below A, in the control and in the presence of 2.8 l M toxin in B (C) The I–V plot of one experiment in which Cn12 was used at concentrations of 0.7 and 2.8 l M in the same cell (D) Plot of the inactivation time constant (s h ) as a function of test potential in the control (h) and from the traces showing the slow toxin-induced component (s) (Materials and methods).Insets: left, control inward current at 0 mV (CON) and in the presence of 2.8 l M Cn12; right, the slowly inactivating toxin-induced current obtained by subtracting the control current from the trace in the presence of toxin.(E) Plot of the voltage-dependent activation and steady-state inactivation (n ¼ 4; Materials and methods).The Boltzmann relationships (dashed lines) were fitted with the following parameters (mV): activation (control) V ½ –10, slope 7.1; inactivation (control) V ½ 41.7, slope 5.2; inactivation (slow component) V ½ 54, slope 6.2 The amplitude of these slowly inactivating traces 3 ms after the onset was plotted (j) as a function of the conditioning potential.The left insets show the currents elicited at 10 mV from )100, )60 and )20 mV in the control (upper) and in the presence of the toxin (lower).The central inset shows the recordings obtained after subtraction of the control traces from the 2.8 l M traces.TTX-sensitive current data are plotted.
Trang 7enough to cause a dramatic change in toxicity to mice
[53,54].In recent years, many publications have
contri-buted novel data and identified possible residues directly
involved in the recognition and binding to Na+channels
[4,5,18,33,55–65].However, a structural and functional
characteristic unique to all ScTXs and Na+channels has
not been found.Rather, each novel toxin needs to be
analysed in the specific context of the structural
determi-nants of its functions and putative receptor or binding sites
For these reasons, the report of a novel function or
structural feature of an undescribed Na-ScTx should be
taken as an important contribution to the field.We have
described here a novel peptide isolated from C noxius
which displays a-like activity.We show that the charge
distribution on the surface of the peptide is probably one of
the significant structural features that govern its function
Furthermore, if we comparatively rotate the 3D structure of
the known Na-ScTxs, it is evident that there are several
equivalent spatial orientations for which a clear difference in
charge distribution exists among these toxins.Here, one
such orientation, called face C, was arbitrarily chosen to
illustrate the point
3D Structural elements of Cn12 There are three secondary-structural elements in Cn12, a two and a half turn a-helix (residues 24–32), three strands
of antiparallel b-sheet (residues 2–4, 37–40 and 45–48), and
a type II turn (residues 41–44) (Fig.4B).These elements are common to all Na-ScTxs.No significant changes were found in the secondary and tertiary structures.It can be seen
in Fig.4C that the segment from residue 5 to 11 and that from residue 53 to 67 is not well defined because of the lack
of long-range NOEs, probably because of the high mobility
of these regions.The a-helix is linked to the b-sheet by two disulfide bridges, which are conserved in all long-chain toxins [8], the CS-a/b motif [2]
Alignment of Fig.2, using the CLUSTAL X program, shows a clear cut separation of all the b Na-ScTxs from all the a Na-ScTxs.The a Na-ScTxs have identities of the order
of 50% among themselves, the same being true for all the
b Na-ScTxs (data not shown).However, when the a and
b Na-ScTxs are compared, the identities fall below 30%.In addition to the eight highly conserved cysteines, Tyr4 and Gly38 are strictly conserved in all toxins, but further
Fig 4 Structure of Cn12 (A) The most representative data for the structure determination of Cn12 (sequential and medium range NOEs, exchangeable amide proton, chemical shift index, and coupling constants) are shown.A strong NOE is represented by a bigger rectangle.Exchange behaviour of amide protons is indicated by black rectangles; the bigger the rectangle the slower the exchange.Arrows indicate b-strands and zig-zag lines indicate the a-helix.(B) Ribbon diagram of Cn12 showing an a-helix at residues 24–32, three strands of antiparallel b-sheet comprising residues 2–4, 37–40 and 45–48, and a type II turn at residues 41–44.(C) The models of 19 out of 250 NMR structures calculated for Cn12 were superimposed, showing well-defined secondary structures for segments of amino-acid residues in positions: 3–4, 24–32, 37–48, except for the N-terminal and C-terminal regions.
Trang 8similarities are present in positions: R1, D2, G3, G10, Y46
and V50 (numbered according to the sequence of Cn12)
Location of charge distribution and the binding affinities
for Na+channels
Data from lethality tests conducted in vivo usually correlate
well with electrophysiological data obtained in vitro A high
toxicity of Na-ScTx usually means high affinity for ion
channels.This seems to be the case for Cn12.It is not toxic
in vivoat concentrations at which other toxins from the
same scorpion, such as Cn2 (a mammalian-specific toxin),
Cn5 (a crustacean-specific toxin) and Cn10 (an
insect-specific toxin) are very effective (doses of 0.4–40 lg per
individual [57,58]).As shown in Fig.3 the affinity of Cn12
for the Na+ channel model chosen for this study (F-11
clone, see Materials and methods) indicates that the affinity
is low (high nanomolar or even 1.0 lM).Thus, it seems to
fit the rule: low toxicity in vivo, low affinity in vitro However, this simple observation could be misleading There are too many variables in the toxin–channel inter-actions.The different tissues of the experimental animals are differently susceptible to different scorpion toxins, as mentioned above.Peptides not toxic when intraperitoneally injected can be highly toxic when intracranially injected [18,22].The types and subtypes of ion channels and other possible receptor targets for the Na-ScTx, and their distribution in cell membranes, are extremely variable and may explain the differences.It is quite clear that scorpions have evolved huge variability in peptides to capture their prey or defend themselves from predators.A plausible explanation for this is the presence of a coevolutionary
Fig 5 Electrostatic surface potentials of selected Na-ScTxs (A) Ribbon diagram of Cn12 (blue) superimposed with Bjxtr-IT (red), Lqh-aIT (green) and BmK M1 (orange), taken from PDB: 1PE4, 1BCG, 1LQH and 1SN1, respectively.This figure was generated after overlapping the solved 3D structures according to the positions occupied by similar secondary-structure elements, as indicated in Fig.2.The orientation chosen corresponds to the original face B described in [66].(B) Electrostatic surface potential of the same toxins calculated with the MOLMOL program [71], using fully charged residues, shown at 1.4 A˚ van der Waals radius.Visible charged residues for each toxin are indicated by the one-letter code, in which red means negatively charged, blue positively charged.Neutral or hydrophobic residues are in white, but not individually marked.Axes in the middle of the figures represent the selected orientation.
Trang 9process.Whenever the channel changes, the toxin also changes in order to most efficiently fit its binding site [64,65].However, as shown in Fig.5A, the Cn12 scaffold
is similar to the others, and yet it is a weak modifier of
Na+channel function
Figure 5B compares the distribution of the charge of Cn12 with three toxins in which site-directed mutations were performed, and the corresponding function studied [18,56,57,63].It is clear that the a-like and a anti-insect toxins (BmK M1 and Lqh-aIT, respectively) show a quite similar overall charge distribution, when analyzed in the orientation originally described as face B [66].Although Cn12 shows an a scorpion toxin effect, it does not have the same charge distribution, suggesting that other faces of the 3D structure of the toxins contribute to this effect.Similarly, the insect-excitatory toxin Bjxtr-IT, defined as b scorpion toxin, has a different charge distribution, as would be expected.The results of site-directed mutagenesis show that face A does not seem to be important for channel recognition.Rather, it probably has a role in maintaining the correct 3D folding of the molecules [55].For example, the residues shown to be important for the function of the
a toxins Lqh-aIT [18] and BmK M1 [55,56] are: K8, Y10, F17, R18, W38, N44, R58, V59 and K8, W38, Y42, K62, H64, respectively.For these two toxins, the residues in question are mostly situated in face B, as shown in Fig.5B However, this seems not to be case for the b toxin Bjxtr-IT
Fig 6 Electrostatic surface potentials for face C of selected Na-ScTxs Same electrostatic surfaces for toxins as in Fig.5B, rotated from face B by 87.5 in the z-axis direction, 62.5 in the y-axis direction, and 10.0 in the x-axis direction, following exactly this order of rotation.This is one of the orientations in which a pronounced difference was found.Toxin structures are labelled as in Fig.5.
Table 1 Experimental constraints and structural statistics.
(A) Distance constraints
(B) Angle constraints 36 (u)
(C) Cartesian coordinate rmsd (A˚) in backbone atoms
Backbone
b-strand and helix 0.612
(D) Energy (kcalÆmol)1) calculated from CNS (19 structures)
Trang 10orientations of the molecule.
On further analysis of the structures available, another
face was found, defined here as face C, which is quite
distinct with respect to charge distribution of the toxins
under analysis.Face C was obtained by rotating the four
superimposed models through 87.5 in the z-axis direction,
62.5 in the y-axis direction, and 10.0 in the x-axis
direction, as shown in Fig.6.Other toxins for which the 3D
structures are known were similarly analyzed: BmK M2,
BmK M4, BmK M8, AaH II, CsE-V (a toxins), and Cn2,
CsE-v5 and Ts1 (b toxins).Using the orientation of face C
for these toxins, a distinct charge distribution is observed
The toxins are like a Christmas tree, decorated in different
forms to interact most efficiently with their specific targets
Comparative analysis of charge distribution in
Na+channels
The receptor site 3 of Na+ channels is surmised to be
formed mainly of extracytoplasmic loops of the S3–S4
segment of domain IV and, to a lesser extent, of the segment
S5–S6 of domain I [14].Contact with a Na-ScTx is mediated
by electrostatic and hydrophobic interactions.Nevertheless,
as shown in Fig 7, Na+-channel isoforms constitute a
highly homogeneous family in terms of primary structure,
particularly in the S3–S4 segment of domain IV.The
variations found among the different isoforms are reduced
remarkable species specificity (e.g Na-ScTx Lqh-II binds
to rat brain synaptosomes with a 100-fold higher affinity than to cockroach preparations; conversely Lqh-aIT binds
to neuronal preparations from cockroach with a 10 000-fold higher affinity than to rat brain synaptosomes [51]).In addition, some toxins are capable of discriminating between
Na+-channel isoforms of the same organism (e.g the rat brain isoform rNaV1.1 is 10-fold more sensitive to the action
of the a Na-ScTx Lqq5 than the cardiac isoform rNaV1.5 [67]).However, when the overall charge of both the toxins and the channels are considered together, a plausible explanation comes from analysis of the other region involved in receptor site 3, segment S5–S6 of domain I.In general, there is greater variability in this segment.Highly sensitive Na+ channels have more acidic residues than insensitive isoforms (compare rNaV1.4 and rNaV1.5 in Fig.7), which would also support the observed preferred interaction of the former with anti-mammal-specific basic
a Na-ScTxs [68–70].In striking contrast, insect-specific a Na-ScTxs usually present an overall neutral or, often, a negative net charge at physiological pH.It has been suggested that the presence of sialic acid in the S5–S6 segment from domain I of mammalian channels may influence the binding of nonpositively charged Na-ScTxs, disfavouring their binding because of the absence of primary electrostatic attraction [70].Cn12 has a slightly negative charge at physiological pH, and the neuroblastoma cells
Fig 7 Multiple alignment of amino-acid sequences corresponding to the receptor site 3 of Na + channels A total of 50 nonredundant sequences of
Na+channels available in databases were aligned with CLUSTAL X [45].The segments S5–S6 of domain I and S3–S4 of domain IV from selected sequences are shown (Na V 1.1, 1.4 and 1.5 from rat; Na V 1.7 from human; Para from fruit fly; NachB1 from squid) Conserved residues are indicated
by asterisks, and conservative replacements by dots.Acidic amino acids in the extracytoplasmic loops are shown in bold.Critical binding residues determined by mutagenesis are highlighted by empty circles [67,70].