E. Wanke, Dipartimento di Biotecnologie e Bioscienze, Università di Milano-Bicocca, Piazza della Scienza, 2U3, 20126 Milan, Italy Fax: +39 02 64483314 Tel: +39 02 64483303 E-mail: email@example.com
α-Conotoxins from marine snails are known to be selective and potent competitive antagonists of nicotinic acetylcholine receptors. Here we describe the purification, structural features and activity of two novel toxins, SrIA and SrIB, isolated from Conus spurius collected in the Yucatan Channel, Mexico. As determined by direct amino acid and cDNA nucleotide sequencing, the toxins are peptides containing 18 amino acid residues with the typical 4/7-type framework but with completely novel sequences. Therefore, their actions (and that of a synthetic analog, [γ15E]SrIB) were compared to those exerted by the α4/7-conotoxin EI from Conus ermineus, used as a control. Their target specificity was evaluated by the patch-clamp technique in mammalian cells expressing α1β1γδ, α4β2 and α3β4 nicotinic acetylcholine receptors. At high concentrations (10 µm), the peptides SrIA, SrIB and [γ15E]SrIB showed weak blocking effects only on α4β2 and α1β1γδ subtypes, but EI also strongly blocked α3β4 receptors. In contrast to this blocking effect, the new peptides and EI showed a remarkable potentiation of α1β1γδ and α4β2 nicotinic acetylcholine receptors if briefly (2–15 s) applied at concentrations several orders of magnitude lower (EC50, 1.78 and 0.37 nm, respectively). These results suggest not only that the novel α-conotoxins and EI can operate as nicotinic acetylcholine receptor inhibitors, but also that they bind both α1β1γδ and α4β2 nicotinic acetylcholine receptors with very high affinity and increase their intrinsic cholinergic response. Their unique properties make them excellent tools for studying the toxin–receptor interaction, as well as models with which to design highly specific therapeutic drugs.
peripheral nervous system nicotinic acetylcholine receptor
central nervous system nicotinic acetylcholine receptor
synthetic α-conotoxin from Conus spurius
nicotinic acetylcholine receptor
α-conotoxin IA from Conus spurius
α-conotoxin IB from Conus spurius
Conotoxins are small, disulfide-rich peptides that have been isolated from Conus, a large genus of predatory marine snails. The primary structures of more than 100 conotoxins have been determined and classified into gene superfamilies on the basis of the amino acid sequences of the signal peptides of their precursors. In general, the members of each superfamily have a characteristic arrangement of their cysteine residues and a particular connectivity of their disulfide bridges. Each gene superfamily comprises one or more pharmacologic families: the O superfamily, containing ω-conotoxins, κ-conotoxins, δ-conotoxins, and µO-conotoxins; the M superfamily, containing µ-conotoxins, ψ-conotoxins, and κM-conotoxins; the S superfamily, containing σ-conotoxins and αS-conotoxins; the T superfamily, containing ε-conotoxins and χ-conotoxins; the P superfamily, containing the spasmodic peptides; the I superfamily, containing several κI-conotoxins, and the A superfamily, containing α-conotoxins, αA-conotoxins and κA-conotoxins .
Competitive antagonists of the nicotinic acetylcholine receptors (nAChRs) belong to the α and αA families. On the basis of the number of residues between the second and third cysteines and on the spacing between the third and fourth cysteines in the mature α-conotoxins, these peptides have been divided into three groups: the α4/7 subfamily, the α3/5 subfamily, and a heterogeneous group including peptides that do not belong to the two previous groups. These groups have different degrees of antagonistic effect on distinct nAChRs: α3/5 toxins block mostly muscular nicotinic acetylcholine receptors α1β1γδ subtypes, whereas α4/7 peptides, with one exception, block neuronal subtypes .
In this article, we describe the purification, amino acid sequence determination and cloning of the cDNA encoding two novel peptides, SrIA and SrIB, found in the venom of Conus spurius. The pattern and the spacing of their cysteines indicate that they belong to the α4/7 subfamily of conotoxins . We also describe a third peptide, [γ15E]SrIB, synthesized by substituting glutamate for the γ-carboxyglutamate residue and used for comparison together with the α-EI conotoxin from Conus ermineus. We showed that results with [γ15E]SrIB were not significantly different from those seen with the natural compounds, and then, owing to the limited amounts of the natural toxins SrIA and SrIB, used mainly this synthetic peptide for long-duration electrophysiologic tests.
The discovery of new agonists or antagonists is of the utmost importance to widen the understanding of alternative functions of nAChRs, which play a crucial role in cellular and molecular mechanisms underlying brain function.
Purification of SrIA and SrIB
Fractionation of C. spurius venom by HPLC, as described in Experimental procedures, gave the profile shown in Fig. 1A. The fractions indicated as SrIA and SrIB were repurified by RP-HPLC, yielding the two pure peptides SrIA and SrIB (Fig. 1B,C), named following the nomenclature proposed by Olivera & Cruz .
Amino acid sequences and cDNA cloning
Automated Edman sequencing of the native peptides SrIA and SrIB unambiguously defined 12 and 13 residues, respectively. Low glutamine signals at positions 12 and 15 of SrIA and at position 15 of SrIB suggested the presence of γ-carboxyglutamate residues at these positions. Residues 3, 4, 9 and 17 of both peptides were tentatively assigned as cysteine (Table 1), on the basis of the absence of any amino acid signal at these positions. This assumption was confirmed directly by the experiments used to determine disulfide bridges (see below). We obtained positive results with PCR amplification of α-conotoxin-type cDNA, reverse transcribed from C. spurius venom duct total mRNA. Two primers known to match the conserved signal peptide-coding region and the 3′-UTR of the α-conotoxin family, respectively , were successfully employed. Exactly the same sequence was obtained from several colonies, which, together with the demonstrated conservation of the signal and propeptide regions, indicated that the amplification protocol was reliable. The deduced SrIA/SrIB precursor sequence agreed with the results of direct peptide sequencing and MS data (see below), and allowed us to define the final unambiguous primary structure for the mature toxins (Fig. 2). From the precursor sequence, and on the basis of earlier observations by our group with toxic peptides , we were also able to predict the amidation of the C-terminal end of the mature toxins. The primary structures of SrIA and SrIB resemble those of previously isolated α-conotoxins with the cysteine framework 4/7 (Table 2).
Table 1. Amino acid sequences and monoisotopic molecular masses of the peptides from C. spurius and of synthetic peptides [γ15E]SrIB and EI.
Experimental mass (Da)
Calculated mass (Da)
Amidated C-terminus; O, hydroxyproline; γ, γ-carboxyglutamate.
The chemical monoisotopic molecular masses of peptides SrIA and SrIB determined by ESI MS are 2202.9 Da and 2158.8 Da, respectively (Table 1). The agreement with the calculated masses (assuming two disulfide bridges and an amidated C-terminus for each peptide, plus one and two γ-carboxyglutamate residues for SrIB and SrIA, respectively) supports the Edman sequence assignment for each peptide. The tentative assignments of amidated C-termini, based on the structure of the precursor (see ‘cDNA cloning’), were confirmed by the ESI MS data.
Determination of disulfide bridges
Two major and more than 20 minor absorbing peaks were observed during the chromatography of peptide SrIA after partial reduction with Tris(2-carboxyethyl) phosphine hydrochloride and alkylation with N-ethylmaleimide (Fig. 3). This high number of derivatives of peptides alkylated with N-ethylmaleimide has been observed in several studies , and it is thought to reflect diastereoisomers resulting from the introduction of a new chiral center in the maleimide ring after formation of the S–C bond during alkylation. Another factor that could generate additional derivatives is the opening of the ring of the N-ethylsuccinimidocysteines by hydrolysis . Selected peptides were sequenced to reveal the positions of the alkylated cysteines. The phenylthiohydantoin (PTH) derivative of N-ethylsuccinimidocysteine elutes between PTH-Pro and PTH-Met in the HPLC system of the sequencer employed. The presence of alkylated cysteines at positions 4 and 17 in some peptides, and at positions 3 and 9 in other peptides, clearly indicated that the connectivity of the two disulfide bridges in peptide SrIA is of the type I–III, II–IV. The absence of peptides with labeled cysteines at positions 3 and 17 or 4 and 9 gives additional support to the proposed disulfide connectivity.
The synthetic peptide [γ15E]SrIB
It has been reported recently that the γ-carboxyglutamic residues present in toxin peptides may be involved in the folding process but are not relevant for their biological activity . Starting from this hypothesis, a peptide sequence was designed that was analogous to those found for SrIA and SrIB, but bearing glutamic acid residues in place of the γ-carboxyglutamic residues at positions 12 and 15 (Table 1). Testing the biological properties of such a peptide, prepared by chemical synthesis and thus with a fully defined chemical structure (including disulfide pattern), would support the amino acid sequence and folding of the native peptides proposed above, and additional tests would not be limited by the availability of the peptide, as might occur with the natural toxins SrIA and SrIB. To obtain the desired folding pattern (see Experimental procedures), we protected the cysteine side chains with two orthogonal protecting groups that can be removed selectively under different conditions, allowing the formation of one disulfide bridge at a time. For this purpose, Cys3 and Cys9 were introduced as S-trityl-protected amino acids, whereas S-(acetamidomethyl)cysteine was used for positions 4 and 17. At the end of chain assembly on the solid support, achieved using standard 2-(1-H-benzotriazol-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate activation protocols for Fmoc chemistry as previously described [9,10], the peptide resin was treated with trifluoroacetic acid for cleavage from the solid support and side chain deprotection, with simultaneous liberation of the two thiol groups in positions 3 and 9. The first disulfide bond was then formed by air oxidation. Finally, the bis-acetamidomethyl-peptide generated was treated with iodine, which caused removal of the protecting group and simultaneous oxidation to disulfide, yielding the fully folded sequence.
Physiologic effects of natural conotoxins and their synthetic analogs
In order to explore the physiologic role of the novel SrIA and SrIB conotoxins, we performed a series of patch-clamp experiments on single cells from the line TE671, which expresses the human muscle receptor , and HEK293 lines stably transfected with the human central nervous system nicotinic acetylcholine α4β2 and peripheral nervous system nicotinic acetylcholine α3β4 receptor subtypes. As our present perfusion system is not sufficiently fast to resolve fast desensitizing currents such as those produced by α7 receptors, we decided not to test our peptides on these receptors, to avoid reporting putatively invalid data. The experiments were done by voltage-clamping the cells at − 60 mV and comparing the responses to brief applications of 50 µm nicotine with those obtained immediately after pretreatment with the different toxins. The concentration of nicotine used during the experiments was fixed at 50 µm, because this value is well below the saturating region of the dose–response curve for the α1β1γδ receptor, as shown in Fig. 4A, and also for the α4β2 and α3β4 receptors [12–14].
The pretreatment time and the concentration of each toxin were varied in the range 3–150 s and 0.2 nm to 10 µm, respectively. A typical experiment performed on a TE671 cell with α-conotoxin [γ15E]SrIB at 1 µm is shown in Fig. 4B. As indicated, the first 50 µm nicotine control pulse produced a response that was strongly reduced after 180 s of toxin perfusion. After 4 min of washout, the application of an additional nicotine pulse produced a recovery that was complete. As the amount of purified toxins was limited, we did the majority of the experiments with the synthetic toxin [γ15E]SrIB and used a known conotoxin , such as EI α-conotoxin, as a control. In the case of the inhibitory effects described in Fig. 4, the results obtained using natural or synthetic peptides, at the same concentration, were not significantly different, and only the data obtained with the synthetic toxin are displayed.
Figure 4C summarizes the data obtained at high toxin concentrations (10 µm). It can be seen that the fractional blockade obtained is both receptor-dependent and toxin-dependent. [γ15E]SrIB was ineffective on α3β4 receptors (n = 4), and was a slightly better blocker of the α4β2 receptors (0.56 ± 0.04, n = 7) than of the α1β1γδ receptors (0.39 ± 0.06, n = 5, not statistically significant). On the other hand, EI toxin was able to potently block muscle (0.95 ± 0.01, n = 5) and ganglionic (0.91 ± 0.03, n = 4) receptors, but was less potent for the central nevous system receptor (0.61 ± 0.02, n = 4). These data not only confirm that EI is an inhibitor of muscle receptors , but they also show that it is a strong inhibitor of the α3β4 receptors and a relatively weak antagonist of the α4β2 central nervous system receptors, on which it had never been tested before.
To further investigate these new EI data, we also performed a series of kinetic experiments at a concentration of 10 µm (see Fig. 4D). EI toxin blocked the α1β1γδ receptors with a τon of 4.9 ± 0.25 s (n = 3), and the α3β4 receptor was blocked with a τon of 11 ± 1.9 s (n = 3). Moreover, the τoff values that we observed for these receptors were 150 ± 13 s (n = 4) and 122 ± 6.5 s, respectively. Figure 4E shows the dose–response curve for the [γ15E]SrIB and EI α-conotoxins on the α1β1γδ receptors. The estimated IC50 and Hill coefficient obtained from these data are: 46 ± 10 nm and 1 ± 0.1 for [γ15E]SrIB, and 187 ± 43 nm and 0.48 ± 0.06 for EI, respectively.
Because, at a toxin concentration [T], a simple Clark's model receptor theory predicts τon = τoff/(1 + [T]/KD), this relationship can be used to confirm the previous IC50 of Martinez et al. on α1β1γδ receptors, which was 280 nm (low-affinity site) for the mouse receptors, and to predict the unknown and novel value of KD for the α3β4 receptors. Indeed, we found an IC50 value of 187 nm for α1β1γδ receptors (Fig. 4E), which also agrees with the fractional response of 0.04 at 10 µm EI and a τoff of 150 s. For the α3β4 receptors, the above relationship results in a KD value of about 1 µm.
On the whole, these experiments, designed to study the antagonistic properties of the toxin [γ15E]SrIB, showed a narrower spectrum of specificity for nAChRs than that of the EI α-conotoxin, owing to the null effect of [γ15E]SrIB on the α3β4 subtype. In contrast, EI was found to be a broad-spectrum α-conotoxin.
During the experiments designed to study inhibitory actions of the two new peptides SrIA and SrIB, we discovered that brief applications, at low toxin concentrations, resulted in increased responses that were immediately reversed after washout of the toxin. A typical experiment performed on an α1β1γδ-expressing cell with different concentrations of α-conotoxin [γ15E]SrIB is shown in Fig. 5A. It can be seen that the first and the last brief control pulses of 50 µm nicotine produced very similar inward currents. However, if 15 s pretreatments with toxin were immediately followed by the same brief nicotine pulses, currents increased, and then decreased as a function of the drug concentration.
In order to shed light on this novel action of the α-conotoxins, we started to investigate whether the various peptides exerted different levels of potentiation on the same α1β1γδ receptor. To clarify whether this novel mechanism was peculiar to the new conotoxins or common also to other, already known, conotoxins, we chose the EI α-conotoxin, which is considered to be an inhibitory conotoxin .
At a fixed toxin concentration of 10 nm, the relative potentiation, (Itoxin − Icontrol)/Icontrol, of the synthetic [γ15E]SrIB, the natural SrIB and SrIA peptides, and the EI α-conotoxin, were as follows: 0.46 ± 0.09 (n = 10), 0.47 ± 0.08 (n = 10), 0.44 ± 0.15 (n = 9), and 0.54 ± 0.13 (n = 9), respectively. These results suggest that, at least for the α1β1γδ receptor type and during brief periods of time (15 s), pretreatment with a concentration of 10 nm toxin shows no clear differences among these peptides. As the amount of natural toxin available for experimentation is limited, and as no significant differences were found when using the synthetic peptide as compared to the native peptides, we continued our assays using the two synthetically prepared products, namely [γ15E]SrIB and EI. The results of these preliminary experiments, obtained only with very low toxin concentrations (0.2 nm to 1 µm) and brief time intervals, do not conflict with those mentioned in Fig. 4, which were obtained with very long pretreatments.
To investigate these mechanisms, the toxins were studied in cells expressing various receptor types. Unexpectedly, we discovered that their effects were also receptor-dependent. To clarify the receptor specificity, we used the two toxins ([γ15E]SrIB and EI) on three different receptors, namely α1β1γδ, α4β2, and α3β4, and the maximally observed relative potentiation values are shown in Fig. 5B. Interestingly, whereas the toxins were unable to produce potentiation in the ganglionic α3β4 receptor (n = 17), the mean fractional potentiation in α1β1γδ receptors for [γ15E]SrIB (0.75 ± 0.22, n = 7) was higher than that obtained for EI (0.35 ± 0.07, n = 22, statistically different). The effects of both toxins were found to be similar on the α4β2 receptor subtype.
Furthermore, we investigated the dose–response curves of the maximal fractional potentiation produced by the [γ15E]SrIB and EI conotoxins on the α1β1γδ receptors. These data are shown in Fig. 5C, and were fitted to dose–response curves with EC50 values of 1.78 ± 1.9 and 0.37 ± 0.23 nm, for [γ15E]SrIB and EI, respectively. An example of this type of action (15 s toxin pretreatment) is shown in Fig. 5D, in one example of an α1β1γδ-expressing cell, with both toxins at two different concentrations (10 and 100 nm). In this experiment, the two toxins were delivered alternately to gain insight into the differences between their sensitivities.
The kinetics of the development of the potentiated response were very fast at concentrations higher than 2–5 nm, and it was almost impossible to determine its time course, given that the rate of bath exchange was < 1 s. However, by reducing the toxin concentration to 0.2 nm, we were able to follow, as a function of the duration of the toxin perfusion, not only the exponential increase in potentiation, but also the decay of the potentiation response, up to the appearance of the blockade. Indeed, if the pretreatment of the toxin lasted for more than 10–15 s, it was possible to observe an exponentially decaying depotentiation process. We show two examples obtained by using the two different toxins on the α1β1γδ receptor. Figure 5E,F shows the potentiation/blockade (Itoxin/Icontrol) data versus duration of toxin pretreatment obtained from experiments done at 0.2 nm[γ15E]SrIB or 0.2 and 1 nm EI, respectively (n = 3). Note the different time scales in Fig. 5E and Fig. 5F. Potentiation data at 1 nm are not shown for [γ15E]SrIB, because they were too fast to be resolved. On the contrary, data at 1 nm for EI, although fast (but not fitted to exponentials), are shown because they illustrate the interesting depotentiation with a time constant different from that observed at 0.2 nm. From these experiments, it can be seen that both the development of potentiation and the depotentiation or block are dependent on the toxin type and concentration. These data suggest a very complex mechanism of toxin–receptor interaction that warrants additional study. Unfortunately, this was beyond the scope of this study.
On the whole, these results suggest that the potentiation described here could be a property of different classes of α-conotoxins. On the other hand, we do not exclude the possibility that this effect could be confined to the conotoxins that act on both neuronal and muscular receptor subtypes, as those used in this work are the only ones reported to be active on both targets. On the α1β1γδ receptor, the synthetic toxin [γ15E]SrIB was less potent than EI, but the latter was less efficient.
Biochemical characterization of SrIA and SrIB
The primary structures of peptides SrIA and SrIB isolated from the worm-hunting snail C. spurius reflect post-translational modifications of proline and glutamine residues, together with the amidation of the C-terminus of a shared toxin precursor. From analysis of the cDNA sequence, the C-terminus, including the last cysteine, is: CGGRR. This sequence is typically present in peptides processed post-translationally. Several rules have emerged from matching the sequences of the mature peptides with the nucleotide sequences of the cDNAs encoding scorpion toxins. If one or two basic residues are present at the C-terminus, they are removed post-translationally. If a glycine precedes the basic residue(s), it is used to amidate the residue preceding the glycine . The MS analyses of toxins SrIA and SrIB showed that these peptides are in fact amidated.
The amino acid sequences indicate that the peptides share structural features typical of the α-conotoxin family. The two peptides contain four and seven residues between the second and the third cysteines, and between the third and the fourth cysteines, respectively (CCX4CX7C). This spacing defines the subfamily of the α4/7-conotoxins (Table 2), the most widespread category of nicotinic antagonists present in cone snail venoms . The α4/7-conotoxins have a conserved proline in loop I, which comprises residues between the second and the third cysteines. Together with Vc1a , peptides SrIA and SrIB are the only known α4/7-conotoxins in which this constant proline is post-translationally modified to hydroxyproline (Table 2). This derivative has been found in µ-conotoxins, ω-conotoxins, κ-conotoxins, κA-conotoxins, αA-conotoxins, ψ-conotoxins, ε-conotoxins, χ-conotoxins, σ-conotoxins, κM-conotoxins, δ-conotoxins, and I-conotoxins . It was also discovered in the α4/7-conotoxin GID , although not at the conserved proline of loop I. Another unusual characteristic of SrIA and SrIB is the presence of γ-carboxyglutamate residues. This post-translational modification has been described in Conus peptides such as the conantokins, the γ-conotoxins, the I-conotoxins, and the ε-conotoxins , and in the N-terminal region of the α4/7-conotoxin GID . However, Vc1a and peptides SrIA and SrIB are the only α-conotoxins in which γ-carboxyglutamate residues occur in loop II, which comprises residues between the third and the fourth cysteines.
Peptides SrIA and SrIB have 18 amino acids and an amidated C-terminus. They are predicted to have charges of 0 and + 1, respectively, at physiologic pH. It has been pointed out that α-conotoxins specific for neuronal subtypes of nAChR are neutral or negatively charged , whereas α-conotoxins that target muscle receptors have a net positive charge . Because, according to these authors, peptide SrIA could be considered a potential antagonist of neuronal nAChR, and toxin SrIB a probable antagonist of muscle nAChR, we decided to test peptides SrIA and SrIB in biological preparations separately expressing neuronal (central, α4β2, and ganglionic, α3β4) and muscle (α1β1γδ) subtypes of nAChR. Unexpectedly, peptides SrIA and SrIB were active on both central and muscle types of the nAChR, which constitutes a novel activity profile of the conserved α4/7-conotoxin-type scaffold. Even more surprising was the finding that peptides SrIA and SrIB have nAChR-potentiating activity, in contrast to all previously studied α4/7-conotoxins.
It has been postulated that divergence within a single superfamily to produce functionally different families is one of the neuropharmacologic strategies employed by the Conus genus, and may account in part for its success in nature .
Structure–function relationship for SrIA, SrIB, and EI
Peptides EI, SrIA and SrIB contain structural elements of the two types of conotoxins that act differentially on neuronal and muscle nAChR. Toxin EI  (present study) and peptides SrIA, SrIB and [γ15E]SrIB are the only conotoxins with a type I cysteine scaffold known to act on muscle nAChR. Except for SrIA, they have positive net charges that might contribute to their activity on muscle receptors , and they (except EI) share with most of the α3/5-conotoxins (blockers of α1β1γδ nAChR) a tyrosine at position 4 of loop II that is not present in any of the α4/7 conotoxins known previously (Table 2). This tyrosine has been found to make an important contribution to the affinity of toxin MI for the α/δ subunit interface of the muscle nAChR . The three peptides have threonines and methionines at position 4 of loop I and position 2 of loop II, respectively. These residues are not present at these positions in any of the other α4/7 toxins studied to date, with the exception of Met10 in toxin EpI (Table 2). It seems probable that these threonines and methionines are somewhat involved in the binding and/or activity with muscle nAChR. Alternatively, the nonpolar methionine residue at position 2 of loop II might be involved in binding to neuronal nAChR subtypes, because all known α4/7-conotoxins have a nonpolar residue at this position (Table 2). Peptides EI, SrIA, SrIB and [γ15E]SrIB have very similar hydrophobic aliphatic residues occupying position 7 of loop II (isoleucine in toxin EI; leucine in peptides SrIA, SrIB, and [γ15E]SrIB); aliphatic residues (leucine, isoleucine, or valine) also occur at this position in toxins MII, PeIA, GIC, Vc1.1, PIA, and GID, which target diverse neuronal subtypes (including α3β4 and α4β2) with variable affinities (Table 2). Thus, it is probable that hydrophobic aliphatic residues at position 7 of loop II contribute to the binding and/or activity of peptides EI, SrIA, SrIB and [γ15E]SrIB with α3β4 and/or α4β2 nAChRs. Finally, except for toxin GID, peptides SrIA, SrIB, and [γ15E]SrIB are the only α4/7-conotoxins known to have an arginine at position 1 of loop II (Table 2). In GID, this residue has been demonstrated to contribute to the block of the α4β2 subtype , which is consistent with the biological activity of peptides SrIA, SrIB and [γ15E]SrIB on α4β2 nAChRs. So far, the toxin with the highest affinity (IC50 = 152 nm) for the α4β2 subtype is GID, and it blocks the α3β2 and α7 subtypes with ∼ 40-fold higher affinities .
The physiologic role of the SrI and EI α-conotoxins
In the present article, we have defined the weak antagonist properties of the novel C. spuriusα-conotoxins, and of a synthetic analog of one of them, on three of the more important types of acetylcholine receptor. Moreover, while comparing these properties with those of the well-known α-conotoxin EI, we discovered that it has a selectivity spectrum somewhat different from that known previously. α-Conotoxin EI had been considered a specific blocker of α1β1γδ nAChRs [15,17,23,24], but our results show that it also may block the α3β4 and α4β2 neuronal subtypes.
This part of our results emphasizes the importance of testing conotoxins not only on the expected subtypes of the known molecular target (based on the toxin sequence and on the current pharmacologic knowledge in the field), but also on other target subtypes and even on nonrelated targets. Recently, toxin ImII has been found to inhibit both α7 and α1β1γδ nAChRs to similar extents , whereas the α3/5-conotoxin CnIA not only inhibits fetal muscle nAChRs, but also blocks the neuronal α7 subtype, although with an ∼ 80-fold lower affinity . One surprising and distinct activity associated with the same protein scaffold of the α4/7-conotoxins has been reported for toxin ρ-TIA from C. tulipa; it inhibits the α1-adrenoreceptor, and has the same disulfide connectivity as ‘classic’α-conotoxins . Like toxin ρ-TIA, which has an extended N-terminal sequence, peptides SrIA and SrIB have sequence features (hydroxylated proline in loop I and γ-carboxyglutamate residues in loop II) that differ considerably from those of other α4/7-conotoxins.
The second part of our results reveals a novel conotoxin-induced functional nAChR state consisting of a potentiation of the response; the potentiation can be detected both with the new toxins and with EI. It can be observed and quantitatively characterized at extremely low concentrations and with brief applications. Interestingly, longer applications produced either a null effect or an inhibitory effect, as expected from the kinetic data shown in Fig. 5E,F and the affinity of the inhibitory process, which were evaluated with prolonged pretreatments.
The α-conotoxins described in this communication showed that they can regulate the nAChR response. It is known  that nAChRs are subjected to a variety of actions, including the increase or decrease of the affinity of the receptor for nicotinic ligands, a phenomenon that may occur in the absence of agonist, and possibly results from stabilization of the desensitized state . Numerous examples of positive and negative allosteric effectors acting at neuronal nAChRs have been reported, illustrating the importance of the allosteric nature of this protein. For example, it was shown that progesterone and 17-β-estradiol act as negative and positive effectors, respectively, of the α4β2 receptor subtype [30,31]. Atropine and zinc are reported to have similar effects on some nAChRs [32,33], although the required concentrations of these drugs were higher by more than two orders of magnitude than those of our peptides. Interestingly, the same mixed partial agonist and antagonist behavior was observed for the well-known blocker d-tubocurarine . It has been reported recently  that α-conotoxin PnIA and a synthetic derivative of it ([A10L]PnIA) weakly potentiate acetylcholine-activated currents in the wild-type α7 nAChR; these authors also reported that on mutant (α7-L247T) receptors, [A10L]PnIA potentiated the acetylcholine-evoked current and acted as an agonist by itself. The mechanisms involved in these processes may be related to previous findings that α-conotoxin MI binds to two distinct sites on the α1β1γδ nAChR, one at the αδ interface, and another at the αγ (or αε) interface .
As it is unknown how and where the peptides studied in this work bind the different receptor assemblies, it is premature to suggest any hypothesis regarding the structure–function mechanisms underlying the peptide binding. Single-channel studies are in progress using mutagenized peptides and cells expressing specific nAChRs.
These peptides are promising tools for studies at a detailed molecular level of the structure–activity relationship that underlies the action of the nAChR-targeting conotoxins. Considering that nAChRs are implicated in brain diseases such as schizophrenia, nocturnal frontal lobe epilepsy , and Alzheimer's disease, these new peptides are also candidate models to develop potentially therapeutic drugs of major importance ; for example, peptides SrIA, SrIB and [γ15E]SrIB might lead to the development of α4β2-selective enhancers, which are beginning to be discovered .
Specimen collection and venom extraction
Specimens of C. spurius were collected in the Yucatan Channel, Mexico. The venom was obtained by dissection of the venom ducts. The ducts were homogenized in 10 mL of 0.1% v/v trifluoroacetic acid and 40% v/v acetonitrile (ACN). The homogenate was centrifuged at 17 000 g for 30 min at 4 °C using a Beckman Coulter Avanti J20 centrifuge with JA-20 rotor. The supernatant, containing the peptides, was subsequently processed.
Peptide purification by RP-HPLC
HPLC was performed on an Agilent 1100 Series LC System (G1322A Degasser, G1311A Quaternary Pump, G1315B Diode Array Detector, G1328A Manual Injector; Hewlett-Packard, Waldbronn, Germany). The venom extract was fractionated with a Vydac (Toluca, Mexico) C18 analytical reverse-phase column (218TP54, 5 µm, 4.6 × 250 mm) equipped with a Vydac C18 guard column (218GK54, 5 µm, 4.6 × 10 mm). Peptides were eluted with a linear gradient of 5–95% solution B at a flow rate of 1 mL·min−1 over 90 min, where solution A is 0.1% v/v aqueous trifluoroacetic acid and solution B is 0.09% v/v trifluoroacetic acid in 90% v/v aqueous ACN. The same column was also employed to repurify the components of the venom, using a linear gradient of 15–30% of solution B at a flow rate of 1 mL·min−1 for 45 min.
Amino acid sequence
Peptides were adsorbed onto polybrene-treated (Biobrene Plus; Applied Biosystems, Foster City, CA) glass fiber filters, and the amino acid sequence was determined by automated Edman degradation using an automatic instrument (Procise 491 Protein Sequencing System; Applied Biosystems) by the pulsed-liquid method.
Native peptides were applied directly into a Finnigan LCQDUO ion trap mass spectrometer (Finnigan, San Jose, CA). The LCQ mass spectrometer is coupled to a Surveyor syringe pump delivery system. The eluate at 20 µL·min−1 was split to allow only 5% of the sample to enter the nanospray source (1.0 µL·min−1). The spray voltage was set to 1.6 kV, and the capillary temperature was set to 130 °C. All spectra were obtained in the positive-ion mode. The acquisition and deconvolution of data were performed with xcalibur software (Thermo Electron Corp., Nashville, TN) on a Windows NT PC system.
Determination of disulfide bridges
The connectivity of the cysteines of toxin SrIA was determined by partial reduction with Tris(2-carboxyethyl) phosphine hydrochloride and alkylation with N-ethylmaleimide. The peptide (11.8 nmol) was dissolved in 10 µL of denaturing buffer (0.1 m sodium citrate containing 6 m guanidine hydrochloride, pH 3.0), and 27 µL of 0.1 m Tris(2-carboxyethyl) phosphine hydrochloride in the same buffer was added. The mixture was incubated for 15 min at room temperature. Sixty-eight microliters of 0.1 mN-ethylmaleimide dissolved in denaturing buffer was added, and the sample was incubated for 30 min at room temperature . The mixture was diluted with 900 µL of 10% solution B, and injected into two, reverse-phase analytical columns in tandem (Sephasil Peptide C8, 5 µm, 4.6 × 100 mm; Pharmacia Biotech, Uppsala, Sweden; Eclipse XDB-C8, 5 µm, 4.6 × 150 mm; Agilent Technologies, Palo Alto, CA) provided with a MetaGuard Nucleosil C8 precolumn (5 mm, 4.6 × 10 mm; MetaChem, Torrance, CA). A linear gradient from 10% to 30% solution B was developed at 1 mL·min−1 for 120 min at room temperature; the absorbance of the eluate was monitored at 220 nm. The absorbing peaks not present in the corresponding reagent blank were collected and taken to dryness for automatic sequence analysis.
The venom duct and gland from one C. spurius specimen was dissected, and total RNA was isolated using the RNAgents Total RNA Isolation System (Promega, Madison, WI), according to the supplier's instructions. Reverse transcription of mRNA was primed by oligo-p(dT)22NN, where p(dT)22 annealed with the polyA tail of mRNA, and NN localized the primer to the border between the polyA tail and the 3′-UTR of the mRNA. Reverse transcription was performed using the 1st Strand cDNA Synthesis Kit for RT-PCR (avian myeloblastosis virus) (Roche Diagnostics Corporation, Indianapolis, IN) as follows: ∼ 1 µg of total RNA and 20 pmol of oligo-p(dT)22NN were added to a volume of 20 µL containing 10 mm Tris/HCl (pH 8.3), 50 mm KCl, 5 mm MgCl2, 1 mm dNTP mix, 50 units of RNase inhibitor, and ∼ 20 units of AMV reverse transcriptase. The reaction was incubated at 25 °C for 10 min, and then at 42 °C for 60 min. The reverse transcriptase was inactivated by incubation at 95 °C for 5 min after the reaction was completed. Conotoxin cDNA was amplified by Vent DNA Polymerase (New England Biolabs, Inc., Beverly, MA), using reverse transcription products as templates, and oligonucleotides corresponding to the conserved signal peptide-coding region and the 3′-UTR of α-conotoxin cDNAs as primers. The primers were: P2, 5′-TCTGCGAATGGGCATGCGGATGATGTT-3′, corresponding to the signal peptide-coding region; and P3, 5′-TGCTCCAACGTCGTGGTTCAGAGGGTC-3′, corresponding to the 3′-UTR . PCR was carried out as follows: to a total 100 µL reaction volume containing 1 × ThermoPol Reaction Buffer [20 mm Tris/HCl (pH 8.8), 10 mm KCl, 10 mm (NH4)2SO4, 2 mm MgSO4, 0.1% Triton X-100] and 0.2 mm dNTP mix was added 5 µL of reverse transcription product and 20 pmol of each primer. The mixture was incubated in the thermal cycler at 95° for 5 min, and then 2 units of Vent DNA Polymerase were added (Hot Start). The parameters for thermal cycling were: 95 °C for 5 min; then 95 °C for 30 s, 47 °C for 30 s, 72 °C for 30 s for 35 cycles; then 72 °C for 5 min (GeneAmp PCR System 2400; Perkin-Elmer, Wellesley, MA). The PCR products were purified with a QIAquick Gel Extraction Kit (Qiagen GmbH, Hilden, Germany), and blunt-end ligated with EcoRV-digested pBluescript II KS (+) vector (Stratagene, La Jolla, CA). The ligation products were electrotransformed into electrocompetent DH5αEscherichia coli cells, and plated on agar plates containing 200 µg·mL−1 5-bromo-4-chloroindoyl-3-yl β-d-galactoside/isopropyl thio-β-d-galactoside/ampicillin to select for recombinant colonies. White colonies were picked, and their plasmids were screened by size on a 1.2% agarose gel . Plasmid DNA was purified with a High Pure Plasmid Isolation Kit (Roche Diagnostics Corporation), and cloned DNA was sequenced with a 3100 DNA Sequencer (Applied Biosystems), using primers T3 and T7, which anneal in the flanking regions of the cloning site of the vector.
Synthesis of peptides [γ15E]SrIB and EI
Solid-phase peptide synthesis
Nα-Fmoc-amino acids, 2-(1-H-benzotriazol-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate and Rink amide MBHA resin were from Novabiochem (Laufelfingen, Switzerland). All other reagents and solvents were from Sigma-Aldrich (St Louis, MO). Mass spectra were recorded on a Fourier Transform Ion Cyclotrone Resonance instrument (model APEXII; Bruker Daltonics, Billerica, MA) equipped with a 4.7 T cryomagnet (Magnex, Oxford, UK). RP-HPLC was performed on a Waters 515 system using a Waters (Milano, Italy) Symmetry 300 C-18 analytical column (5 µm, 4.6 × 150 mm) or a Merck (Darmstadt, Germany) LiChroCART 250 C-18 semipreparative column (10 µm, 10 × 250 mm). Linear gradients of ACN in water/0.1% trifluoroacetic acid were used for peptide elution, with flow rates of 1 mL·min−1 for analytical or 8 mL·min−1 for semipreparative purposes. The peptides were prepared by manual, solid-phase peptide synthesis using standard 2-(1-H-benzotriazol-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate activation protocols for Fmoc chemistry as previously described [9,10]. Couplings were carried out with a five-fold excess of activated amino acid for a minimum of 30 min, and monitored by the 2,4,6-trinitrobenzenesulfonic acid test . At the end of chain assembly, the resin was washed with dimethylformamide/dichloromethane, dried under nitrogen, and treated with a mixture of cation scavengers in trifluoroacetic acid with magnet stirring for 2 h to detach the peptides from the resin and simultaneously to deprotect amino acid side chains. The crude peptides were precipitated in hexane/diethylether 1 : 1 (v/v), centrifuged at 3290 g using a Hereus Sepathec centrifuge with 2251 rotor, dissolved in 20% aqueous ACN, and purified by semipreparative HPLC [43,44].
The S-acetamidomethyl (Acm)-protected precursors of the [γ15E]SpIB and EI peptides had the following sequences: RTCC(Acm)SROTCRMEYPELC(Acm)G-NH2, and RDOCC(Acm)YHPTCNMSNPQIC(Acm)-NH2, respectively.
Mass analysis of purified peptides gave the following results: peptide RTCC(Acm)SROTCRMEYPELC(Acm)G-NH2, MS 2262.0 (calculated 2262.5); peptide RDOCC(Acm)YHPTCNMSNPQIC(Acm)-NH2, MS 2237.3 (calculated 2237.8).
Formation of the first disulfide bond
The lyophilized peptides were dissolved in 10 mm aqueous ammonium bicarbonate to a final peptide concentration of 5 mm. The pH was adjusted to 8 by adding 1.5 m Tris/HCl buffer (pH 8.8). The resulting solution was stirred for 48 h and then concentrated. The bis-Acm-peptides containing disulfide bridges between Cys3 and Cys9 in the case of [γ15E]SrIB or between Cys4 and Cys10 in the case of EI were purified by semipreparative HPLC.
Mass analysis gave the following results: Bis-Acm-[γ15E]SrIB precursor, MS 2260.6 (calculated 2260.8); Bis-Acm-EI precursor, MS 2235.1 (calculated 2235.4).
Acm group removal and formation of the second disulfide bond
The bis-Acm peptides were dissolved in 80% aqueous acetic acid to a final peptide concentration of 1 mm. Three equivalents (eq.) of 37% hydrochloric acid were added. Iodine (10 eq., methanolic solution) was added dropwise, and the resulting mixture was stirred for 2 h. The reaction was then quenched by adding 1 m ascorbic acid until the orange color disappeared. The mixture was concentrated, and the final product was isolated by HPLC and lyophilized. This procedure resulted in the formation of disulfide bridges between Cys4 and Cys17 in [γ15E]SrIB and between Cys5 and Cys18 in EI. Mass analysis gave the following results: [γ15E]SrIB, MS 2114.8 (calculated 2115.0); EI, MS 2075.4 (calulated 2075.8).
Cells of the rhabdomyosarcoma TE671 clone  were routinely cultured in DMEM containing 4.5 g·L−1 of glucose and 10% fetal bovine serum. The cells were incubated at 37 °C in a humidified atmosphere with 5% CO2. Cells expressing α3β4 nAChRs  were kindly provided by D. Feuerbach (Novartis Pharma AG, Basel, Switzerland). Cells expressing α4β2 nAChRs were kindly provided by G. Casari (S. Raffaele Institute, Milan, Italy).
The standard extracellular solution contained: 130 mm NaCl, 5 mm KCl, 2 mm CaCl2, 2 mm MgCl2, 10 mm Hepes/NaOH buffer, and 5 mm d-glucose (pH 7.40). The standard pipette solution at [Ca2+]i = 10−9m (pCa 9) contained: 130 mm K+-aspartate, 10 mm NaCl, 2 mm MgCl2, 0.2 mm CaCl2, 10 mm EGTA/KOH, 10 mm Hepes/KOH buffer, and 1 mm ATP (Mg2+ salt) (pH 7.30). Fresh nicotine and conotoxins were added every 2 h to the extracellular medium from a stock solution in distilled water.
Currents were recorded on an MC700A patch-clamp amplifier (Axon Instruments, Union City, CA) at room temperature, as previously described . Pipette resistance (1–2 MΩ), cell capacitance and series resistance errors were carefully compensated (85–95%) before each voltage-clamp protocol was run. The extracellular solutions were ejected through a nine-hole (0.6 mm) remote-controlled linear positioner (average response of 0.5–1 s) placed near the cell under study or by the similar rotating perfusion system (Biologic, Grenoble, France). Tygon rather than silicon tubing (Tygon-R3603; Cole-Palmer, Vernon Hills, IL) was used for extracellular solutions, to prevent nicotine adsorption. We used nicotine instead of acetylcholine to exclude the action of the latter on muscarinic receptors that are potentially present on the cell membrane, and we verified that the responses described in the TE671 cells could be blocked by 500 nmα-bungarotoxin (not shown). The flow rate was constant at 120 µL·min−1. The holding potential was set at − 60 mV. Data acquisition and analysis were performed with pclamp 8.2 (Axon Instruments) and origin 7.1 (Microcal Software, Inc., Northampton, MA) software, respectively.
We thank Maria Eugenia Ramos for the collection of C. spurius specimens, Patricia Villalobos and Andrés Falcón for technical assistance, Leopoldo Martínez for artwork, and Dorothy D. Pless for revision of the manuscript. This work was supported by grants from Ministero dell'Università e della Ricerca Scientifica e Tecnologica (MURST-COFIN 2001-03 and 2003-05; FISR-Neurobiotecnologie: Fisiopatologia del sistema nervoso; FIRB-2001) to E. Wanke, Consejo Nacional de Ciencia y Tecnología (CONACYT) to E. P. Heimer de la Cotera (41477-Q) and M. B. Aguilar (43754-Q), and Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica-UNAM (PAPIIT-UNAM) to E. P. Heimer de la Cotera (IN-204403). R. Restano Cassulini and E. Schiavon are PhD students in Physiology at Milano-Bicocca University, Department of Biotechnology and Biosciences.