Transgenic Tarantula Toxin: A novel tool to study mechanosensitive ion channels in Drosophila

Graphical abstract


Introduction
Venom from the tarantula Grammostola spatulata contains a potent and specific inhibitor of mechanosensitive ion channels (MSCs). The active agent is a short 34 amino acid peptide called Grammostola spatulata mechanotoxin #4 (GsMTx4) (Suchyna, et al., 2000). It is the only known toxin specific for cation-selective MSCs, with known inhibitory activities for the channels of the Piezo and Transient Receptor Potential (TRP) families (Bae et al., 2011;Spassova et al., 2006;Alcaino et al., 2017). It has become an important tool to identify and characterise the physiological roles of MSCs (Bowman et al., 2007;Suchyna, 2017).
GsMTx4 is commercially available and is easily applied to ex vivo preparations adding to the power of the toxin as a tool. However, there are some contexts where it would be desirable to deliver the toxin in a non-invasive way. For example, when studying development, integrated physiological systems or behaviour. Furthermore, for some studies it would be advantageous to deliver the toxin to specific organs, tissues or cells. This would be difficult or impossible when the only method for delivery is exogenous application. For these reasons we generated transgenic fly lines with a genetically encoded GsMTx4 gene under the control of the UAS-GAL4 system to allow spatial and temporal delivery of the toxin in the intact animal. Here we describe these lines and assess their ability to inhibit MSCs in the context of mechanical nociception in the larval peripheral nervous system.

Results & discussion:
We synthesized the Grammostola spatulata gene encoding the GsMTx4 peptide, cloned it downstream to the UAS inducible promoter and generated transgenic fly lines (Fig. 1). Two versions were generated: (i) a full-length variant containing the full GsMTx4 cDNA (GsMTx4-FL) encoding a pre-pro-peptide including a N-terminal signal peptide and a pro-sequence, both of which are subsequently cleaved to make the active peptide; (ii) a variant consisting of the active peptide alone (GsMTx4-AP). As GsMTx4-AP lacks the signal peptide sequence it will not be secreted, and would restrict MSC inhibition to the cells in which it is expressed.
We tested the function of the transgenic GsMTx4 lines by assessing their ability to block mechanical nociception in larval pickpocket-GAL4 (ppk-GAL4) sensory neurons -a subset of multidendritic neurons that tile the body wall and respond to a variety of external stimuli including mechanical force ( Fig. 2A) (Zhong et al., 2010). We first checked to determine whether GsMTx4-FL or GsMTx4-AP affect ppk-GAL4 neuron viability or morphology, and find they do not (Fig. 2B, C). Painful mechanical stimuli are transduced by ppk neurons to elicit an escape behaviour in which the larva contracts into a crescent then rolls vigorously with corkscrew-like revolutions away from the stimulus ( Fig. 2D and Supplementary Movie 1). We find mechanical stimulation with a 50 mN von Frey filament elicits the escape response in > 90% of T wild-type 3rd instar larvae (Fig. 2E, Supplementary Movie 1). A number of MSCs, including the Piezo channel, are known to be required in ppk-GAL4 neurons to transduce nociceptive stimuli; perturbation of any of these channels attenuates escape behavior (Zhong et al., 2010;Kim et al., 2012;Tracey et al., 2003). We used Piezo KO larvae as a positive control to score for deficits in mechanical nociception. In line with a previous study (Kim et al., 2012), we find a significant proportion (30%) of Piezo KO larvae do not respond to the 50 mN stimulus (Fig. 2E). Next, we tested the ability of GsMTx4 to block mechanical nociception by driving expression in ppk-GAL4 neurons. For this experiment we used ppk-GAL4 heterozygote larvae as controls. We find > 75% ppk-GAL4 heterozygotes respond to the stimulus (Fig. 2E). In contrast, expression of GsMTx4 in ppk-GAL4 neurons severely attenuates the escape response, with only 46% (GsMTx4-FL) and 35% (GsMTx4-AP) of larvae responding to the stimulus (Fig. 2E, Supplementary Movie 2). These data show that genetically encoded GsMTx4 efficiently blocks mechanical nociception in ppk-GAL4 neurons when expressed in vivo. The phenotype for GsMTx4-AP is significantly stronger than GsMTx4-FL (P = 0.02, Fig. 2E). As GsMTx4-AP is designed not to be secreted it is likely the toxin reaches higher concentration within the neurons, possibly explaining its greater potency. We find expression of GsMTx4 (FL and AP) in ppk-GAL4 neurons does not result in any other obvious physiological or behavioural defects in either larvae, pupae or adults. In future experiments, it will be interesting to assess the effects of in vivo expression of GsMTx4 in other contexts where MSCs have been implicated, for example for epithelial morphogenesis in tissues during embryonic dorsal closure (Hunter et al., 2014), or for stretch-activated mechanotransduction in bipolar dendritic sensory neurons (Suslak, et al., 2015). Our results show the genetically encoded GsMTx4 efficiently blocks mechanical nociception in ppk-GAL4 neurons and strongly imply that its potency as an inhibitor of MSCs is maintained when expressed in vivo. A further implication of our findings is that the toxin is active from the intracellular face of the membrane in line with some, but not all, previous studies (Bae et al., 2011;Nishizawa and Nishizawa, 2007;Kamaraju et al., 2010). An important question that remains is which channel(s) are targeted by GsMTx4? Our data show ppk-GAL4 > GsMTx4 (FL and AP) produce stronger phenotypes than Piezo KO (P ≥ 0.0001), suggesting GsMTx4 inhibits other channels. Three MSCs (Piezo, Painless and Ppk) are expressed in ppk-GAL4 neurons and are required for mechanical nociception (Zhong et al., 2010;Kim et al., 2012;Tracey et al., 2003). It is conceivable the GsMTx4 phenotype is a combined phenotype resulting from inhibition of all three channels. There is strong evidence for GsMTx4 inhibition of channels from Piezo (although the site of inhibitory action appears to be the extracellular face of the membrane) and TRP (Painless is a TRP channel) families (Bae et al., 2011;Spassova et al., 2006;Alcaino et al., 2017). For this reason, we suggest Piezo and Painless are possible targets of GsMTx4 in ppk-GAL4 neurons. However, the phenotype observed with GsMTx4 cannot be explained simply by the combined inhibition of Piezoand Painless (even if both are targeted) because genetic data indicate they are part of the same pathway (the phenotype of the double homozygote is no stronger than the homozygote phenotype of either mutation alone) (Kim et al., 2012). It is therefore necessary to postulate additional targets. Pickpocket (an ENac/DEG channel) is a candidate given that it is present and required in ppk-GAL4 neurons for mechanical nociception, although there is no evidence yet to indicate that GsMTx4 is able to inhibit ENac/DEG channels. A further candidate is the recently described Piezo-like channel (Hu et al., 2019), however it is not known whether Piezo-like is required for mechanical nociception or if it is expressed in ppk-GAL4 neurons. GsMTx4 has also been shown to influence the activity of the TREK mechano-gated potassium channels (Gottlieb and Sachs, 2009;Gnanasambandam et al., 2017). Therefore, it is possible that the loss of mechanical nociception reported here could be accounted for by the activation of the Drosophila TREK channel (Sandman) as well as inhibition of MSCs.
It is conceivable that GsMTx4 inhibits multiple different classes of MSCs. Supporting this, there is strong evidence that GsMTx4 exerts its inhibitory effects indirectly by modulating local mechanical tension in the vicinity of the channel upon insertion between membrane lipids, rather than through direct interaction with a specific channel (Suchyna et al., 2004). Future studies combining genetic mutations, GsMTx4 expression and electrophysiological recordings would resolve the identify of MSCs targeted by GsMTx4.
The use of a genetically-encoded tetanus toxin light chain in Drosophila has proved to be a powerful tool to identify physiological functions of neurons and to map neuronal circuits (Sweeney et al., 1995). Likewise, we suggest the lines described here will provide a powerful new tool to study MSCs and mechanotransduction in vivo. They allow the delivery of the toxin to specific organs, tissues or cells, and can be adapted to provide temporal control of expression simply by crossing into appropriate genetic backgrounds (McGuire et al., 2004). The lines will be particularly useful for studying mechanotransduction during development, physiology and behaviour in intact, free moving animals.

Nociception assay
Mechanical nociception was tested as described previously using a calibrated von Frey filament (Kim et al., 2012). Briefly, wandering 3rd instar larvae were transferred into a petri dish with shallow water and allowed to acclimatise for at least 2 min. The mechanical stimulus was delivered at a 90° angle from above to the dorsal side of abdomen (in the region of segments 5-7). A positive response was scored if at least one 360° rotation was elicited by the stimulus. Each larva was tested only once. The von Frey filament was constructed using 0.2 mm diameter nylon monofilament fishing line attached to a cocktail stick (de Sousa et al., 2014). Filament force (mN) was calculated using a laboratory balance by measuring the mass (g) upon filament bending and multiplying by the gravitational acceleration constant (g; 9.8). The 12 mm filament used in this study produced a force of 49.78 mN ± 0.43 (mean ± s.e.m, n = 15).

Larval imaging
A 3rd instar larva was immobilised beneath a coverslip on a microscope slide. Coverslip bridges either side of the larva prevented D. Beqja, et al. Journal of Insect Physiology 127 (2020) 104116 crushing. GFP was imaged live using a Nikon A1R FLIM confocal mi-