Philanthotoxin Analogues That Selectively Inhibit Ganglionic Nicotinic Acetylcholine Receptors with Exceptional Potency.

Philanthotoxin-433 (PhTX-433) is an active component of the venom from the Egyptian digger wasp, Philanthus triangulum. PhTX-433 nonselectively inhibits several excitatory ligand-gated ion channels, and we recently showed that its synthetic analogue, PhTX-343, exhibits strong selectivity for neuronal over muscle-type nicotinic acetylcholine receptors (nAChRs). Here, we examined the action of 17 analogues of PhTX-343 against ganglionic (α3β4) and brain (α4β2) nAChRs expressed in Xenopus oocytes by using a two-electrode voltage clamp at -100 mV. IC50 values for PhTX-343 inhibition of α3β4 and α4β2 receptors were 7.7 and 80 nM, respectively. All the studied analogues had significantly higher potency at α3β4 nAChRs with IC50 values as low as 0.16 nM and with up to 91-fold selectivity for α3β4 over α4β2 receptors. We conclude that PhTX-343 analogues displaying both a saturated ring and an aromatic moiety in the hydrophobic headgroup of the molecule demonstrate exceptional potency and selectivity for α3β4 nAChRs.


Introduction
A constituent of the venom of the Egyptian digger wasp, Philanthus triangulum, is the polyamine-containing toxin, known as philanthotoxin-433 (PhTX-433; Fig. 1), which enables paralysis of its insect prey through inhibition of ionotropic glutamate receptors (iGluRs) and nicotinic acetylcholine receptors (nAChRs). 1,2 The structure of PhTX-433 consists of a central tyrosine residue amide-linked to a thermospermine moiety on one side and to an n-butanoyl chain on the other (Fig. 1). The resulting molecule thus has a relatively bulky and hydrophobic 'headgroup' and a positively charged (+3) 'tail' at physiological pH. Although the wasp has evolved to produce toxins targeting iGluRs and nAChRs in insects, PhTX-433 and its structurally very similar synthetic analogue, PhTX-343 (1; Fig. 1), are also potent inhibitors of vertebrate ionotropic receptors. These latter interactions have been well characterized at mammalian iGluRs, including the N-methyl-D-aspartate (NMDA), kainate and α-amino-3hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors; 3,4 as well as for vertebrate muscle-type 5,6 and neuronal-type 7,8 nAChRs. In our recent study on neuronal nAChRs, it was found that PhTX-343 is a very potent inhibitor of heteromeric receptors with selectivity for the ganglionic 34 subtype. 7 (1), used as a template for the analogues studied in the present work.
Inhibition of both nAChRs and iGluRs by PhTX-343 has been shown to be use-and voltagedependent (with increased inhibition occurring at more negative membrane potentials), and this has led to the hypothesis that open-channel blocking is the dominant mode of action, with the polyamine tail penetrating deep into the channel pore and interacting with polar amino acids, while the headgroup interacts with rings of hydrophobic amino acids at a more extracellular location in the pore. 9,10 Further evidence for this mechanism resulted from experiments showing that AMPA receptors lacking the GluA2 subunit are highly sensitive to PhTX-343, whereas those containing GluA2 are almost insensitive. 11 This can be explained by RNA editing of GluA2, resulting in a single amino acid substitution at the so-called "Q/R site" that forms the selectivity filter within the pore. 9 Likewise, mutations in the outer pore of nAChR affect the affinity for PhTX-343. 7 Studies into the importance of the amine functionalities resulted in the first breakthrough in receptor selectivity. Thus, the analogue PhTX-12, in which the two secondary amines in PhTX-343 were replaced with methylene groups, had an expected reduced potency at AMPA receptors and NMDA receptors, but unexpectedly inhibition of muscle-type nAChRs was significantly increased. 4 However, the latter was accompanied by a change in the mode of action, since inhibition remained strong at positive membrane potentials and voltage dependence was weak. 6,10 In another study it was found that exchange of the tyrosine moiety in PhTX-343 for a cyclohexylalanine unit generated a toxin with considerably enhanced inhibitory activity at muscle-type nAChR. 12 In the present work, we investigated PhTX-343 (1) and its headgroup-modified analogues (i.e., compounds 2-18; Table 1) with single or combined structural alterations in the n-butanoyl, tyrosine and polyamine moieties for their inhibitory actions against the α4β2 and α3β4 subtypes of neuronal nAChRs. We aimed to develop analogues capable of distinguishing between neuronal heteromeric nAChR subtypes. Recently, there has been considerable interest in compounds that selectively inhibit the α3β4 subtype of nAChRs as a potential treatment for nicotine addiction due to their significant expression in the medial habenula that modulates the mesolimbic dopaminergic response to nicotine. 13 Hence, potent and selective α3β4 nAChR antagonists have the potential to aid smoking cessation. 14, 15

Potency of analogues with substitution of tyrosine only
Three of the investigated analogues (2-4) contained alternative central amino acids instead of tyrosine (Tyr), thereby conferring increased hydrophobicity to the side chain (Table 1).
Compounds 2 and 3 with a phenylalanine (Phe) moiety or its homologue (hPhe) instead of Tyr exhibited increased potency at both nAChR subtypes, which was more pronounced for α3β4, thus leading to increased selectivities of 30-fold and 12-fold, respectively. Compound 4, having a cyclohexylalanine (Cha) residue in place of Tyr, proved more potent than 1 at both nAChR subtypes, but noticeably all selectivity for α3β4 over α4β2 was abrogated ( Fig. 2; Table 1).
Interestingly, analogue 4 was the only compound tested in this study that showed this nonselective behaviour, while it was the most potent inhibitor of α4β2 receptors (Fig. 2B).

Potency of analogues with alternative N-acyl moieties
In analogues 5-6, the N-acyl moieties comprise cyclohexanoyl or cyclopentanoyl instead of the original N-butanoyl group, which confer increased bulk and hydrophobicity to the headgroup (Table 1). Analogues 5 and 6 showed a strong increase in potency at α3β4 receptors, but less so at α4β2 receptors, and hence selectivity for α3β4 was increased up to 80-fold for compound

Potency of analogues with modifications in the polyamine tail moiety
Analogues 7 and 8 contain a cyclopropane ring in the centre of the polyamine moiety in a trans or cis conformation, respectively (Table 1). Both compounds exhibited moderately increased potency at α3β4 receptors as compared to that of 1, whereas they only displayed weakly improved inhibition of α4β2 receptors. This gave rise to increased selectivity of 7 and 8 for α3β4 receptors to 48-fold and 33-fold, respectively.  Table 1.

Potency of analogues with a combination of two headgroup modifications
This set of ten compounds all contain a Tyr  Cha substitution. The effect of this change alone can be seen in the inhibitory profile of analogue 4, for which no selectivity for α3β4 over α4β2 was found. Six of the compounds (i.e., 9-14), containing small to large saturated ring systems instead of the n-butanoyl moiety (Table 1), showed minor to moderately increased potency at α3β4 receptors, while displaying reduced to moderately increased potency at α4β2 receptors, except for analogue 11 that was a 15-fold more potent inhibitor than 1 at the latter receptor.
Selectivity for α3β4 was variable within this series of six compounds with 11 having slightly reduced selectivity as compared to that of 1, while 9 and 13 proved to be considerably more selective (43-fold and 51-fold, respectively) than the original lead compound (i.e., 1).
The remaining four compounds in this subclass (i.e., 15-18) all have aromatic N-acyl substituents ( Table 1). Except for 18, containing an extended and rigid biphenyl group, the introduction of aromatic moieties resulted in very strongly increased inhibitory potency (as compared to that of 1) at α3β4 receptors. By contrast, only weak to moderate enhancement of potency at α4β2 was observed for these analogues, resulting in a highly improved selectivity for α3β4. In fact, compound 16, carrying a naphthoyl group proved to be the most potent and selective among all the analogues tested in the present study, as it was 48-fold more potent than 1 at α3β4 (i.e., with an IC50 = 0.16 nM) combined with a selectivity of 91-fold in favour of α3β4 ( Fig. 2 and 3).

Discussion and Conclusion
In the present work it was demonstrated that analogues of PhTX-343 (1) can be designed to exhibit extremely high antagonistic potency at nAChR receptors, and in particular at the α3β4 subtype. The most potent analogue was 16, containing an N-naphthoyl-Cha headgroup and a spermine polyamine tail, with an IC50 of 0.16 nM. This is certainly the most potent known PhTX-343-derived antagonist of nAChR, and to our knowledge it is the most potent nAChR antagonist reported to date. Additionally, compound 16 has high selectivity for α3β4 over α4β2, and based on our recent observations that PhTX-343 selectively inhibits 4-containing nAChRs (especially α3β4) over all of the other major nAChR subtypes (i.e., α4β2, α3β2, α7 and α1β1γδ) 7 it is expected that this will be the case also for 16. Many of the compounds described in the present work also compare favourably to a recently identified synthetic αconotoxin, TP-2212-59, that selectively targets α3β4 nAChRs with a low nanomolar IC50 value (2.3 nM). 18 However, a direct comparison between compounds belonging to these two distinct classes of antagonists are not straightforward due to their different binding sites and mode of action.
The proposed binding site for PhTX-343 and its analogues, having similar highly charged polyamine tail regions, is located within the nAChR pore, whereby the polyamine moiety inserts deeply into the pore in order to interact with the threonine and serine rings beyond the equatorial leucine gate, while the bulky and hydrophobic headgroup interacts with hydrophobic residues of the valine and leucine rings that line the outer part of the pore. 10 Compound 16 possesses modifications in both the N-acyl and the central amino acid residue, and it may thus be considered to be a further modification of PhTX(Cha)-343 (4), which previously was identified as an antagonist of human muscle nAChRs in TE671 cells, where it proved 277-fold more potent than 1 at VH = -100 mV. 12 Nevertheless, in the present study, by using the same holding potential, only a 2.9-fold increase in relative potency of 4 (versus that of 1) at α3β4 was found, most likely arising from the fact that 1 is much more potent (317-fold) at α3β4 than at the α1β1γδ subtype. 7 Interestingly, the four most potent analogues at α3β4 receptors, having IC50 values < 0.5 nM might also be expected to be a highly potent and selective inhibitor of α3β4 receptors, however, its aromatic moiety is quite extended and rigid, which may restrict access to its binding position due to steric hindrance.
Other highly potent compounds comprise 11, 13 and 14 that have a Tyr  Cha substitution and a bulky saturated hydrophobic aliphatic ring structure instead of the linear n-butanoyl moiety. This increases headgroup hydrophobicity considerably, and it is beneficial for interaction with valine and leucine residues in the outer pore regions of both nAChR subtypes.
Furthermore, compounds 2 and 7 also had sub-nanomolar IC50 values, and 2 was previously found to have increased potency at Torpedo electric organ nAChRs through testing in a [ 3 H]H12-HTX binding assay, 19 presumably because of an increased headgroup hydrophobicity induced by the Tyr  Phe substitution resulting in the absence of a hydroxyl group. Compound 7 has a cyclopropane functionality inserted into the centre of the spermine moiety, and previous modelling studies indicate that internal hydrogen bonding between the amine closest to the headgroup and the amide oxygens will position the cyclopropane unit so that it contributes to the bulkiness of the headgroup. 10,20 At α4β2 receptors analogue 4 was the most potent antagonist, whereas it was one of the least potent compounds at α3β4 receptors. This most likely arises from the absence of an aromatic moiety in the headgroup, providing improved hydrophobic interactions with the valine and leucine residues in the outer pore region of α4β2 than any of the other compounds. The second most potent compound (i.e., 11) at α4β2 also lacked an aromatic group, however, other nonaromatic compounds had lower potency, presumably due to lack of optimal shape and/or size restrictions.
In summary, PhTX-343 analogues displaying both a saturated ring structure and an aromatic moiety in the headgroup demonstrate extreme potency and selectivity for the α3β4 nAChR subtype, and it is hypothesised that this arises from the unique presence of a Phe residue in the outer pore region of this receptor subtype. The most selective compounds identified may have potential as smoking cessation therapies and find application as useful probes in the study of nicotine addiction.

Experimental Section
Materials. Starting materials (amino acid building blocks, trityl chloride polystyrene resin, and reagents) and solvents were obtained from commercial suppliers (Iris Biotech, Sigma-Aldrich, VWR, Labscan and CHEMsolute) and used as received. CH2Cl2 was distilled from P2O5 and stored over 4 Å molecular sieves.
Acetylcholine (ACh) was from Sigma. PhTX-343 (1) was synthesized as described in Wellendorph et al. 21  were prepared in a similar manner to that reported previously. 16 In brief, trityl chloride polystyrene resin (1.66 g, 2.2 mmol/g) in a syringe (20 mL, fitted with a polypropylene filter and a Teflon valve), was swelled in dry CH2Cl2 and then treated with 10% iPr2EtNCH2Cl2 Following wash with DMF (5 ×), a preincubated (for 15 min) mixture of the appropriate carboxylic acid (5 equiv), HOBt (5 equiv), and DIC (5 equiv) in dry DMF (3 mL) was added to the respective reactors, which then were shaken at room temperature for 16 h. The resins were then drained and washed with DMF, MeOH, DMF, and CH2Cl2 (each 3 ×). Cleavage of the products from the resin was performed with 50% TFACH2Cl2 (5 mL for 2 h). The resins were further eluted with CH2Cl2 and MeOH (each 2 × 5 mL); the total eluates from each reactor were co-evaporated with tolueneMeOH 1:1 (3 × 10 mL). The resulting crude products were dried overnight, and then purified by preparative HPLC (by using a linear gradient rising from 5% B to 100% B during 20 min) to give PhTX analogues 5, 6, 10, 11, and 13-17 as the tris(TFA) salts. Compounds 10 and 11 required rechromatography (using a linear gradient from 5% B to 60% B during 20 min) to obtain samples with sufficient purity.    ACh at the end of the one-minute application (i.e., "late" current) in order to allow for equilibration of the inhibitory effect, and hence to achieve improved comparison of the effect of test compounds. Currents in the presence of each concentration of each PhTX analogue were normalised as percentage of the control response (i.e., to ACh alone), and subsequently these were plotted against PhTX concentration. Graphpad Prism 7 was used for data analysis, graph plotting, and curve fitting. All plotted data points are the mean ± SEM, obtained from analysis

Notes
The authors declare no competing financial interest.