The α1-adrenoceptor inhibitor ρ-TIA facilitates net hunting in piscivorous Conus tulipa

Cone snails use separately evolved venoms for prey capture and defence. While most use a harpoon for prey capture, the Gastridium clade that includes the well-studied Conus geographus and Conus tulipa, have developed a net hunting strategy to catch fish. This unique feeding behaviour requires secretion of “nirvana cabal” peptides to dampen the escape response of targeted fish allowing for their capture directly by mouth. However, the active components of the nirvana cabal remain poorly defined. In this study, we evaluated the behavioural effects of likely nirvana cabal peptides on the teleost model, Danio rerio (zebrafish). Surprisingly, the conantokins (NMDA receptor antagonists) and/or conopressins (vasopressin receptor agonists and antagonists) found in C. geographus and C. tulipa venom failed to produce a nirvana cabal-like effect in zebrafish. In contrast, low concentrations of the non-competitive adrenoceptor antagonist ρ-TIA found in C. tulipa venom (EC50 = 190 nM) dramatically reduced the escape response of zebrafish larvae when added directly to aquarium water. ρ-TIA inhibited the zebrafish α1-adrenoceptor, confirming ρ-TIA has the potential to reverse the known stimulating effects of norepinephrine on fish behaviour. ρ-TIA may act alone and not as part of a cabal, since it did not synergise with conopressins and/or conantokins. This study highlights the importance of using ecologically relevant animal behaviour models to decipher the complex neurobiology underlying the prey capture and defensive strategies of cone snails.

ρ-TIA into fish did not reveal a phenotype 16 , previous studies examining the effects of prazosin on fish suggest ρ-TIA might also induce a sleep-like state in fish 18 . However, with the exception of the conoinsulins 19 , the behavioural effects of these potential nirvana peptides added to aquarium water housing fish has not been established to confirm their potential role in net hunting.
In this study, we analysed behavioural effects on the teleost zebrafish induced by synthetic candidate nirvana cabal conotoxins added to their surrounding water to characterize their potential to contribute to net hunting. Zebrafish larvae behaviour was monitored using an automatic tracking system (Zebrabox Revolution) that allows real-time measurement of swim speed, scoot distance and angle of turn behaviours [20][21][22][23][24][25][26][27] . Surprisingly, conopressins and conantokins had no detectable effect on the fish 'fight or flight' response either alone or in combination, whereas ρ-TIA potently blocked the zebrafish escape response to mechanical touch stimuli. N-terminal truncations of ρ-TIA and site-directed mutagenesis of the zebrafish α 1 -AR confirmed that ρ-TIA acted at a related allosteric site in the zebrafish α 1 -AR. This study directly implicates ρ-TIA as an antagonist at the zebrafish α 1 -AR that may contribute to the nirvana cabal, allowing fish capture directly by mouth without harpooning.

Systemic effects of dissected C. tulipa venom in adult zebrafish.
To determine the venom duct localisation of peptides contributing to harpoon prey capture, we extracted C. tulipa dissected venom from four duct sections, proximal (P), proximal central (PC), distal central (DC) and distal (D), and administered 0.1 µg of each intramuscularly (i.m.) and monitored for any behavioural changes. Venom from each of the duct sections reduced the swimming ability of fish, with the P section being most potent and causing an 80% reduction in total swim distance relative to the control (Fig. 1A). Dissected venom from each section produced flaccid paralysis that drastically slowed swimming movements, consistent with a motor cabal effect. Fish injected i.m. with the P dissected venom displayed an immediate and almost complete loss of motor activity that was irreversible over 15 min. In contrast, fish administered with the DC dissected venom had a delayed onset of activity, while PC and D had similar but weaker effects, reducing the total swim distance by 50% relative to the control.
Dissected C. tulipa venom did not induce a nirvana cabal effect in larval zebrafish. In an attempt to establish the duct localisation of nirvana cabal peptides, we administered dissected venoms to the fish water column and monitored for any change in fish behaviour using 5-day post fertilisation (5-dpf) zebrafish larvae in multi-well assay plates. Addition of dissected C. tulipa duct venom (0.001-100 ng/µl) from each of the duct sections failed to alter the fish escape response (Fig. 1B) and all larvae displayed normal swimming behaviour.
Conantokin and conopressin have no effect on fish escape responses. Administration of dissected venom to larval water failed to produce nirvana cabal-like activity in zebrafish larvae, perhaps because the relevant peptides are secreted only on demand. We thus decided to test the hypothesised synthetic nirvana peptides individually and in combination on larval zebrafish. Neither the VR antagonist conopressin-T nor the NMDAR antagonist conantokin-T from C. tulipa affected the larval escape response at concentrations up to 1 µM (Fig. 2). Similarly, neither the VR agonist conopressin-G nor conantokin-G from C. geographus affected the larval escape response at concentrations up to 5 µM (Fig. 2). In all treatment groups, exposed larvae displayed normal swimming behaviour indistinguishable from the control group. ρ-TIA reduces the fish escape response in larval zebrafish. Following a preliminary screen of synthetic conotoxins, we identified that ρ-TIA (0.03-10 µM) added to the water column produced a striking loss of zebrafish larvae escape response to light mechanical touch on their trunk or tail. This effect was dose-dependent, with an EC 50 of 0.19 ± 0.62 µM. ρ-TIA at concentrations above 3 µM produced an acute reduction in swimming speed, with the larvae exhibiting an apparent rigid body with minimal gill movements and abnormal escape responses (escape response score 2-3 vs. a control score of 8.9 of 10). At 1 µM, reduced escape responses developed over the initial 5 min of exposure to ρ-TIA, allowing larvae to be easily pushed around with the Teflon prod without showing escape or avoidance responses. (Fig. 3, Supplementary Video). The effects of 1 μM ρ-TIA lasted beyond the 15 min observation period, while prolonged exposure to ρ-TIA (>3 μM) was lethal to larvae. Additionally, ρ-TIA induced bursts of erratic swimming, similar to a stress response but with a stronger swimming response (Supplementary Fig S1). This altered swimming behaviour was observed for ~1 min after addition of ρ-TIA at concentrations above 0.01 μM, and was most pronounced at 0.3 µM ρ-TIA.
Assessing synergistic effects of ρ-tiA with presumed nirvana cabal conotoxins. To determine if the observed nirvana cabal effect results from synergy between different peptide classes, we treated the larvae with different combinations of conopressins, conantokins and ρ-TIA, each dosed at 1 µM. Co-exposure of zebrafish to conopressins and conantokins from C. geographus (Fig. 4A) or C. tulipa (Fig. 4B) had no effect on larval zebrafish escape responses, and exposed zebrafish instantly displayed the characteristic "C-start" 28 response when presented with the mechanical stimulus. The escape responses for ρ-TIA in the presence of the conantokin/conopressin combinations were also indistinguishable from the responses observed for ρ-TIA alone (Fig. 4).

Discussion
The venom of piscivorous snails has been extensively studied for its pharmacological potential and more recently to understand the evolution of their predatory and defensive strategies. The Gastridium clade, including the well-studied C. geographus and C. tulipa, have developed a net hunting strategy to catch teleost fish directly by mouth using venom secreted directly into the surrounding water. However, the conotoxins contributing to this unique net hunting strategy remain unclear. In this study, we developed a zebrafish larval assay using an automated screening platform to quantify behavioural changes. Surprisingly, the dissected C. tulipa venom from each of the four duct sections added to the water column had no effect on the zebrafish escape response, suggesting nirvana cabal peptides are normally present in the venom duct at low concentrations and presumably upregulated and secreted into the water during net hunting. In contrast, i.m. injection of these dissected venoms produced flaccid paralysis in zebrafish that dramatically reduced their ability to swim, especially the proximal duct venom where defensive venom peptides are produced in the closely related C. geographus 2 , supporting their putative role in defence in C. tulipa.
Previous studies have suggested conantokins, conopressins, contulakins and conoinsulins as the putative nirvana peptides 11,19,29 . However, apart from the conoinsulins, the putative role of these non-paralytic conotoxins is based on their effects on rodents. In the search for putative nirvana cabal peptides, we used the zebrafish research model to assess the potential of conopressins and conantokins, commonly found in C. geographus and C. tulipa venom, to contribute to the nirvana cabal. Contrary to expectation, conopressins and conantokins previously identified from C. geographus and C. tulipa, administered either alone or in combination to aquarium water, had no significant effect on zebrafish escape responses and overall behaviour. In contrast, the α 1 -AR receptor antagonist ρ-TIA added to aquarium water dramatically reduced the ability of the zebrafish to respond to mechanical stimuli. ρ-TIA was previously found to be non-lethal to fish by injection 16 and was prominent in the central but not the proximal duct section 15 , suggesting it does not contribute to the defence-evoked venom of C. tulipa. Co-addition of conopressins and conantokins together with ρ-TIA failed to alter the ρ-TIA effect, suggesting that conopressins and conantokins probably do not contribute to a nirvana cabal in C. tulipa and raising the possibility that ρ-TIA may act alone to facilitate net feeding in this species. It remains to be determined if the conopressins and conantokins contribute to the nirvana cabal in C. geographus together with the contulakins and conoinsulins commonly found in this species.
ρ-TIA has previously been shown to be specific for α 1 -ARs over ligand and voltage-gated ion channels in nerves and skeletal muscle 16 . To identify the target of ρ-TIA in zebrafish, we assessed the ability of N-terminal truncations of ρ-TIA, previously identified as critical for allosteric inhibition at mammalian α 1 -ARs 30 , to inhibit the zebrafish escape response. In parallel with reductions observed in mammalian ARs, these analogues had a reduced potency to inhibit the escape response, consistent with inhibition of fish α 1 -ARs being the target for the observed sensory deprivation, and suggesting that ρ-TIA acts similarly on mammalian and zebrafish α 1 -ARs. ρ-TIA did not fully inhibit 3 H-prazosin binding to zebrafish α 1 -ARs, consistent with the non-competitive effect seen for mammalian α 1 -ARs 31 . The introduction of mammalian α 1 -AR pharmacophore residues previously identified as important for ρ-TIA binding 17 into the zebrafish AR (E327D/S330F-ZFα 1BB -AR) enhanced ρ-TIA affinity 36-fold. Together, these data indicate that ρ-TIA can act at low nanomolar concentrations and likely binds at the same allosteric site on mammalian and zebrafish α 1 -ARs. Interestingly, both of the mutated pharmacophore residues are found in ZFα 1AA -and ZFα 1AB -ARs (Table 2), while F330 is found in ZFα 1BA -AR, suggesting ρ-TIA may be a useful probe to help identify which of the four zebrafish α 1 -ARs contribute to this loss of escape response observed.
In addition to inhibiting the escape response of fish, ρ-TIA treated larvae also showed erratic bursts of swimming, reminiscent of the hyperactivity described for excitatory κA-conotoxins in adult zebrafish 32 . However, the excitatory effects observed with ρ-TIA lasted ~1 min and were fully reversible, with larvae resuming normal swimming after the hyperactive state, in contrast to the irreversible paralysis produced by κA-conotoxins. Neurotransmitter function is an important regulator of locomotion in adult fish, and disruption by antipsychotic drugs such as fluphenazine and haloperidol produces similar erratic swimming motion in zebrafish 33 . It is possible that upon initial exposure, ρ-TIA may temporarily interfere with the lateral line sensory neurons, disrupting normal function.
Through a combination of behavioural and pharmacological assays on zebrafish, we have identified that ρ-TIA present in C. tulipa venom retains its antagonistic properties towards the fish α 1 -ARs and reverses the stimulatory  effects of norepinephrine on fish to induce a "sleep-like" quiescent state 18 . This observed biological effect appears to be parallel to the sensory deprivation effect induced during net predation, which might facilitate fish capture by mouth by these nocturnal feeders. Thus, we hypothesise that ρ-TIA plays a direct biological role in net hunting by C. tulipa. The low levels of ρ-TIA identified in the venom duct suggest that secretion of ρ-TIA is stimulated only when C. tulipa starts net hunting. In contrast, conopressins and conantokins appear to play little role in the net hunting strategy of C. tulipa, raising questions about their biological function in this venom. Conantokins in particular have a dominant distal venom expression 13,15 and yet there is no definitive study implicating their potential role in prey capture and/or defence. Given conantokins are expressed in several piscivorous venoms [34][35][36][37] , their involvement in the hook and line prey capture strategy warrants further investigation. Similarly, the inactivity of conopressin-T in the behavioural assays further supports its potential role in the defensive venom of C. tulipa 15 .
Although ρ-TIA may act alone to facilitate net hunting in C. tulipa, we cannot exclude the possibility that other conotoxins, yet to be identified, may act in synergy. In contrast, net hunting in C. geographus may be facilitated by conoinsulins, which alone have been shown to affect fish behaviour at 25 µM 19 , and/or contulakins, which are both prominent in C. geographus but not C. tulipa venom. Finally, development of a milking strategy for the net hunters of the Gastridium clade that allows direct isolation and sequencing of the "nirvana cabal" proteome is required to further characterise the peptides responsible for this unique predatory strategy that profoundly alters fish behaviour. Such peptides are expected to provide new research tools that are useful in the elucidation of the neurobiology of escape responses and sleep in fish. Venom peptides. A single specimen of C. tulipa was sacrificed and the venom duct was dissected. The duct was divided into four equal parts; proximal (P), proximal central (PC), distal central (DC) and distal (D) section and dissected venom extracted separately from each section, as described previously 38 . Conantokin-G, conantokin-T, ρ-TIA and its N-terminal truncated analogues were chemically synthesized as described below. All peptides were made up to the required stock concentrations in E3 medium and stored at 4°C prior to use.

Simulation of the nirvana cabal.
To evaluate the effect of venom/peptides on the behaviour and motor function of the zebrafish larvae, we analysed both the response of the animals to a manual mechanical touch response assay and their spontaneous swimming behaviour using the Zebrabox Revolution automatic tracking/screening platform (ViewPoint, France) 21 . Briefly, 5-dpf larvae and adult fish (500-600 mg) were exposed to dissected venom (0.001-100 ng/μl), ρ-TIA and its analogues (0.001-10 μM), conantokins (0.003-1 μM) and conopressins (0.003-5 μM). The larvae and adults were tested in a 24-multiwell plate and a 0.5 L tank (Tecniplast, USA) respectively. The plates/tank were incubated in the Zebrabox Revolution and the animals were automatically tracked with a user-defined programme. Experiments were performed in light conditions and the swimming distance was defined as small (speed of 4 cm/s) and large (speed of 8 cm/s) distances. Zebrafish were monitored for a 5 min pre-treatment period and a 10 min post-treatment period, and data binned in 2 s intervals. Raw data and video were exported and processed in silico to export total swim distance for each group. At 15 min, treated fish were subjected to the touch response assay, as described below.
touch response assay. Larvae escape response was evaluated using a manual mechanical touch using a www.nature.com/scientificreports www.nature.com/scientificreports/ 1-10 (1, absence of movement; 2, fin vibration without movement; 3, one body coil without movement; 4, several body coils with slight movement; 5-9, gradual increment in swim distance; and 10 normal escape response with larvae reaching edge of the well). For each concentration, six larvae were tested individually, with each test comprising 10 separate manual stimulus scores. Untreated negative control responses (E3 medium only) were recorded in parallel using the same protocol.
Systemic effects of conotoxins. Adult fish were administered 0.1 μg of dissected venom from each of the four venom duct sections by intramuscular (i.m.) injections (5-10 μL) using a Hamilton syringe (cemented point style 2, 26 s gauge needle) (Sigma-Aldrich no. 20795-U). Injected animals were placed in 1 L tanks under ambient light conditions, and monitored using the Zebrabox revolution (Viewpoint) imaging platform. Baseline swimming readings were obtained for 15 min before administering 5 μL of the dissected venom and the resulting behavior was recorded for further 15 min. Zebrafish were analysed for alterations in swimming pattern, and swimming distance relative to baseline quantified for each fish. clone design and site-directed mutagenesis. The ZF α 1BB -AR cDNA (NCBI Reference Sequence: NM_001007358.2) was designed and cloned into pcDNA3.1 (Genescript, USA), whereas the hamster α 1B -AR cDNA was cloned into the pMT2' vector (kind gift from Professor Bob Graham; Victor Chang Cardiac Research Institute, Sydney, Australia). The ZF α 1BB -AR cDNA was subjected to in vitro site-directed mutagenesis to generate the double mutant E327D/S330F α 1BB -AR cDNA, using specific primer pairs (Sigma Aldrich, Australia) and following the QuikChange mutagenesis kit (Stratagene). TOP10 Escherichia coli (Invitrogen) were transformed with WT or mutant cDNA and subsequently used for plasmid preparation using a PureLink ® Quick Plasmid Miniprep Kit (Invitrogen) or High Speed Maxi kit (Qiagen). Purified cDNA was used to confirm all mutations by sequencing by the Australian Genome Research Facility. transient expression of α 1 -ARs and membrane preparation. COS-1 cells (ATCC) were cultured in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) containing glutamine and 5% fetal bovine serum. Cells were transiently transfected with purified plasmid DNA encoding WT or mutant α 1 -ARs using FuGENE HD (Promega) following the manufacturer's protocol. 48 h post-transfection, cells were harvested and homogenized using a Polytron homogenizer (Brinkmann Instruments) in HEM buffer (20 mM HEPES, 1.5 mM EGTA, 12.5 mM MgCl 2 , pH 7.4) containing complete protease inhibitor (Roche Diagnostics). The homogenate was centrifuged at 484 × g (2000 rpm) for 10 min and the resulting supernatant was centrifuged at 23,665 × g (14000 rpm) for 30 min. The pellet was resuspended in HEM buffer containing 10% v/v glycerol and stored at −80°C prior to use. Protein concentration was determined using the BCA protein assay kit (Pierce) following the manufacturer's protocol.
ip-one HtRf functional assay. COS-1 cells (ATCC) were cultured and transiently transfected with WT or mutant α 1 -AR DNA using FuGENE HD (Promega) following the manufacturer's protocol. Assays measuring IP 1 -accumulation were performed 48 h post-transfection following the manufacturer's instructions (IP-one HTRF assay kit, Cisbio international). In brief, the effect of increasing concentrations of norepinephrine (NE) (100 pM-100 µM), or ρ-TIA (100 pM-100 µM) were determined in the presence of 10 µM NE for WT and mutant ZF α 1BB -AR or 1 µM NE for WT hamster α 1B -AR. The assay was performed on 30,000 transfected cells in stimulation buffer in white 384 multiwell plates (Optiplate, PerkinElmer Life Sciences) incubated for 1 h at 37°C with 5% CO 2 . Cells were then lysed by the addition of HTRF reagents before europium cryptate-labeled anti-IP 1 antibody and the d2-labeled IP 1 analog were added (IP-one HTRF assay kit, Cisbio International), followed by incubation for 1 h at room temperature. Emissions at 590 nm and 665 nm were recorded after excitation at 340 nm using a microplate reader (Tecan). Statistical analysis. Data were analysed using GraphPad Prism software, v7.0a (San Diego, USA). In vivo dose-response data was fitted to non-linear regression curves to establish EC 50 s. IP-one data was expressed as the HTRF ratio F = [(fluorescence 665nm /fluorescence 590nm ) × 10 4 ] normalized to the response elicited without ligand (negative control). For both the IP-one and the radioligand binding assay, non-linear regression curves were used to establish EC 50 and IC 50 values. All IC 50 and EC 50 values are expressed as the mean ± SEM averaged from six independent experiments, unless stated otherwise.