The VAChTY49N mutation provides insecticide-resistance but perturbs evoked cholinergic neurotransmission in Drosophila

Global agriculture and the control of insect disease vectors have developed with a heavy reliance on insecticides. The increasing incidence of resistance, for virtually all insecticides, threatens both food supply and effective control of insect borne disease. CASPP ((5-chloro-1’-[(E)-3-(4-chlorophenyl)allyl]spiro[indoline-3,4’-piperidine]-1-yl}-(2-chloro-4-pyridyl)methanone)) compounds are a potential new class of neuroactive insecticide specifically targeting the Vesicular Acetylcholine Transporter (VAChT). Resistance to CASPP, under laboratory conditions, has been reported following either up-regulation of wildtype VAChT expression or the presence of a specific point mutation (VAChTY49N). However, the underlying mechanism of CASPP-resistance, together with the consequence to insect viability of achieving resistance, is unknown. In this study, we use electrophysiological characterisation of cholinergic release at Drosophila larval interneuron→motoneuron synapses to investigate the physiological implications of these two identified modes of CASPP resistance. We show that both VAChT up-regulation or the expression of VAChTY49N increases miniature (mini) release frequency. Mini frequency appears deterministic of CASPP activity. However, maintenance of SV release is not indicative of resistance in all cases. This is evidenced through expression of syntaxin or complexin mutants (sytx3-61/cpxSH1) that show similarly high mini release frequency but are not resistant to CASPP. The VAChTY49N mutation additionally disrupts action potential-evoked cholinergic release and fictive locomotor patterning through depletion of releasable synaptic vesicles. This observation suggests a functional trade-off for this point mutation, which is not seen when wildtype VAChT is up-regulated.


Introduction
Vesicular transporters load neurotransmitter into synaptic vesicles (SV) for storage before release. Transporter localisation dictates loading substrate and differs between vesicle classification [1]. Small clear SVs mostly store fast-acting neurotransmitters. Transporters known to PLOS [2][3][4], the glycine transporter VGLYT1 [5,6], the vesicular GABA and Glycine co-transporter VGAT/ VIAAT [7][8][9], VAChT [10] and vesicular monoamine transporters (VMATs) [11]. The insect CNS and to a lesser, but still significant, extent the mammalian CNS relies on ACh and VAChT for excitatory synaptic signalling. Drosophila VAChT null mutants die during embryogenesis [12] whilst mammalian VAChT null mutants are lethal soon after birth [13]. VAChT knockdown also results in memory deficits in both mice and insects [14,15]. By contrast, VAChT up-regulation causes accelerated neuromuscular aging [16] and disrupted central cholinergic morphological development [17]. Altered physiological and morphological phenotypes are also reported in Drosophila at the glutamatergic neuromuscular junction (NMJ) following VGLUT up-regulation [18]. It is, perhaps, unsurprising that transporter abnormalities are associated with many neurodegenerative diseases including Alzheimer's [19], Huntington's [20] and Parkinson's disease [21]. Cholinergic central interneurons form a central pattern generator that drives Drosophila motoneurons and are pivotal in governing the control of larval peristalsis [22]. Normal cholinergic function is essential for survival and, as a result, the cholinergic system is a major target for insecticide development [23]. CASPP compounds specifically bind to, and inhibit the function of, VAChT [24][25][26][27][28][29]. CASPP mortality is associated with an inability of central neurons to release ACh [30]. However, transgenic up-regulation of VAChT significantly reduces CASPP lethality in adult Drosophila, to an extent greater than observed in C. elegans [25] and, moreover, greatly increases spontaneous mini release of ACh from larval Drosophila central premotor interneurons [30]. This suggests a causal relationship between up-regulation of target activity and resistance to CASPP. Furthermore, transgenic expression of the VAChT Y49N mutation, in an otherwise wildtype background, provides complete insensitivity to CASPP [25]. However, the physiological implications to CNS function of the presence of this resistance allele, and its response to insecticide treatment, have not been described.
Insecticide resistant phenotypes have been formally identified since the discovery of DDT resistance in Musca domestica [31] and have been characterised into four sub categories; (i) behavioural changes, (ii) altered penetration, (iii) target site modification and (iv) metabolic resistance [32]. Through these mechanisms, insects have attained resistant alleles for the majority of pesticide classes including but not limited to organophosphates, carbamates, pyrethroids, neonicotinoids and non-neuronal insecticides such as insect growth regulators (IGRs) [33][34][35][36]. However, little is known about the physiological implications brought about by resistant genotypes. Here we utilise Drosophila as a model to study CASPP resistance generated through transgenic up-regulation of wildtype VAChT transcript achieved via the GAL4/UAS system and through the endogenous expression of a CRISPR-induced knock-in of VAChT Y49N . Electrophysiological characterisation of cholinergic minis in identified Drosophila motoneurons (specifically the anterior corner cell, aCC and Raw Prawn 2, RP2 motoneurons) shows that CASPP resistance is associated with the maintenance of mini release frequency. However, despite trends in enhanced synaptic transmission, we further demonstrate that similarly increased mini release, achieved through up-regulation of mutant membrane bound SNARE protein syntaxin (sytx  [37] or the absence of complexin (cpx SH1 ) [38], does not lead to CASPP-resistance. This suggests that resistance to this class of insecticide cannot be fully explained by enhanced mini release at cholinergic synapses. Perturbation of evoked release (i.e. action-potential dependent) associated with VAChT Y49N additionally disrupts the ability of the cholinergic locomotor system to maintain sustained activity attributed by an apparent change in SV release. These deficits translate into fitness costs observed through abnormalities in locomotion and adult longevity. any additional role in data collection and analysis or decision to publish, The specific roles of these authors are articulated in the 'author contributions' section.

Competing interests:
The authors have the following competing interests: Dr. Goodchild is employed by Syngenta Crop Protection. The insecticidal activity of the 5-CL-CASPP disclosed in this paper has been patented by Syngenta (WO2003106457, WO2005058897, WO2005061512), however the focus of this study (the molecular mechanism of action) is independent of any commercial interest, and there are no other relevant Syngenta patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials. SWV and RAB declare no competing interests.

gRNA and insert design, template oligo and plasmid construction
The CRISPR Optimal Target Finder tool (http://tools.flycrispr.molbio.wisc.edu/targetFinder/) was used to specify target cut sequence specificity (GATTACCGCTATCAGGTACC). Two guide RNA constructs were made to generate cuts in 5'-and 3'-UTR of VAChT, respectively. The gRNA oligonucleotides (5' to 3') are: 5'-UTR: CTTCGAGAGGAAGTCCCAAAGAAAC and AAACGTTTCTTTGGGACTTCCTCTC; 3'-UTR: CTTCGTATTATTACTATAGACATAT and AAACATATGTCTATAGTAATAATAC, sense and antisense, respectively). 100 pmol of each 5' phosphorylated sense and antisense gRNA oligonucleotides were mixed, denatured at 95˚C and then reduced to 25˚C at a rate of -0.1˚C/sec and ligated to the guide RNA expression plasmid, pU6-BbsI-chiRNA (plasmid #45946, addgene). A VAChT Y49N , containing 5' and 3' PAM site mutations, that prevent cas9 cleavage, was cloned to pHD-DsRed vector (plasmid #51434, addgene) as a dsDNA donor template for CRISPR/Cas9-mediated homology-directed repair (HDR). Oligos used to generate PAM site mutations are depicted in Table 1. Briefly, for 5' PAM site mutagenesis, PCR of primers a+b and c+d (containing TGG to TGC point mutations) were run against Drosophila genomic DNA (PCR1). Following purification, PCR products (a+b and c+d) were used as templates for a second PCR using the most 5' and 3' primers of PCR1 (primers a and d, Table 1). This process was repeated for 3' PAM site mutagenesis utilising primers (e+f and g+h). Full UTR sequence with PAM mutations were purified, sequenced and mobilised to pHD-DsRed using restriction digests (5' = AscI & BssSI), 3' = SpeI & XhoI). VAChT Y49N ORF was generously provided by Syngenta and cloned into pHD-DsRed vector using restriction enzymes (EcoR1 and Nde1). Sequence confirmation was confirmed by Sanger sequencing at the Manchester Sequencing Facility.

Identification of positive progeny
Positive progeny were identified by the expression of DsRed in larvae following the 3xP3 expression pattern. Single adult males were selected and individually balanced using CyO and TM3. Fully balanced lines were sequenced for the VAChT Y49N sequence. DsRed was removed using a Cre-recombinase line (y 1 w 67c23 P {Crey} 1b; D Ã /TM3, Sb 1 , flybase ID: FBst0000851). Lines were then homozygosed using (w -; if/CyO; MKRS/TM3) and cleaned generating a final stock of w -; +; VAChT Y49N .

Mortality and 5-Cl-CASPP application
5-Cl-CASPP (also termed Syn351), made by Syngenta [25], was added to food vials containing standard cornmeal medium. 5-Cl-CASPP was solubilized in acetone (which was added alone as vehicle control). CASPP solution (50μl) was added to each vial and blunt forceps were used to break up the food surface to aid penetrance. A total of 60 first instar larvae were assayed per drug concentration, 20 larvae per vial. Larvae were reared at 25˚C for 7 days (in a humid incubator). Mortality rate was measured as the proportion of larvae developing to pupae, against those that did not. For use in electrophysiology assays, 5-Cl-CASPP (10μg/ml) was added to the surface of grape-agar plates (50 mm diameter) in 1 ml of aqueous dried yeast extract (5%; Merck, Darmstadt, Germany). After being left to dry overnight at room temperature, second instar larvae were allowed to feed for 24 h before recording.
Evoked vesicle exocytosis, mimicking action potential mediated spontaneous rhythmic currents (SRCs), were elicited through driving UAS-ChR2 Cheta (λ470 nm, 10ms, 1Hz/10Hz) using the cholinergic driver ChAT-BAC in the background of the wild type VAChT or mutant VAChT Y49N . Stimulations were carried out for up to 1 hour. As is true for other cholinergic drivers (e.g. cha B19 -GAL4), ChAT-BAC shows weak expression in aCC (our unpublished observations). Recordings conducted in the presence of the nAChR antagonist, mecamylamine, reveal a small~20pA inward depolarising current in unison with stimulating light pulses (i.e. due to this weak expression). Therefore, events 20pA, measured using the Clampfit threshold search function, were considered a failed event and thus the point at which the vesicle pool had been depleted.

Larval extracellular recordings
Bursts of spiking driven by cholinergic synaptic input in aCC/RP2 motoneurons were investigated using loose patch extracellular recordings. Recordings were conducted using a 1.3MO recording electrode filled with extracellular solution allowing a loose seal on the soma when applying gentle suction. Recordings, sampled at 20 KHz, were made for 3 minutes with the second minute of each recording taken for analysis. The number of activity bursts within this second minute of recording, together with the number of spikes per burst (for the first 10 bursts) were quantified and averaged for each recording. Bursts were defined as a minimum of three events within 25 ms.

Adult giant fibre activation
Flies were sedated on ice for 5 mins before being mounted in dental wax (Kemdent, UK). Appendages, wings and proboscis were secured into the wax with thorax being positioned at a 45˚angle aiding penetration of recording electrodes. Two sharpened tungsten wires (0.2mm) were inserted into the CNS through each compound eye whereas the third was positioned into the abdomen. Stimulation was conducted with a SIU5A stimulus isolation unit (Grass Technologies, USA) controlled via a S88 stimulator (Grass Technologies, USA). Positioning of both stimulating and earth electrodes was confirmed with a pulse (50V, 0.02 ms) through observation of wing twitching (DLM stimulation) and tergotrochanteral muscle contraction (TTM stimulation). Intracellular recording electrodes (20-30 MO) (GC100FS-10; Harvard Apparatus, WPI, USA) were filled with 3M KCl for recording. Glass recording electrodes were inserted into DLM 45a and the TTM, identified by position of the bristles on the thorax. Recordings were performed using an Axoclamp-2A amplifier (Axon instruments, USA) controlled by pClamp 10.4 and a Digidata 1440A (Molecular Devices, USA). The responsiveness of the respective muscle (DLM and TTM) to a train of 10 stimuli at 100Hz, termed "Following Frequency" were recorded.

Larval crawling
Single larvae were placed on a 55 x 34mm, 2% nutrient free agarose island surrounded by 5M NaCl to prevent escape. Recordings were conducted for 3 minutes using EthoVision XT (Version 11, Noldus, Wayeningen, Netherlands) in a Daniovision Behavioural Chamber (Noldus, Wayeningen, Netherlands), using a Basler GenIcam (Basler acA1300-60, Resolution: 1280x960, Frame Rate: 15Hz). Background was adjusted for using the inbuilt dynamic subtraction function, and a smoothing function of 10 was applied (each data point was a function of 10 data points averaged). Larvae were considered moving when breaching a 0.30mm distance threshold. Recordings were discharged if sample detection threshold was 15% of total recording. Mean velocity (mm/s) and total distance covered (mm) within each 3 minute recording were taken for analysis.

Oviposition and longevity
Number of eggs laid per mated female was taken as one measure of overall fitness. 10-20 virgin females were allowed to mate for 48 hours under standard rearing conditions before being transferred to egg laying chambers mounted on a grape-agar plate (50 mm) supplemented with yeast paste. Following a 24 hour period, total number of embryos per plate was quantified.
Following oviposition, single mated females from each genotype were added to single vials and allowed to lay for 5 days. The presence of viable larvae was considered proof the female had been mated. Total life span was further taken as a measure of general fitness. Newly eclosed male flies were added to standard rearing vials at a density of 10 flies per vial per replicate. Flies were transferred to fresh medium three times per week and deaths recorded, as per [42].

Quantitative RT-PCR (qRT-PCR)
Ten CNSs were collected from third-instar wall-climbing larvae. RNA was extracted using the RNeasy micro kit (QIAGEN, Manchester, UK). Single strand cDNA was synthesized using the Revert Aid™ H minus First strand cDNA synthesis kit (Fermentas, Massachusetts, USA). qRT-PCR was performed using a LightCycler480 II (Roche, Basel, Switzerland) with SYBR Green I Master reaction mix (Roche, Basel, Switzerland). The thermal profile used was 10 min at 95˚C followed by 45 cycles of 10 s at 95˚C, followed by 10 s at 60˚C, and finally 10 s at 72˚C. Single-product amplification was confirmed by post-reaction dissociation analysis. PCR primers were designed with the aid of LightCycler Probe Design Software 2.0 (v1.0) (Roche, Basel, Switzerland). Results were analysed by the 2 -ΔCt method where ΔCt was determined by subtracting the average rp49 Ct value from that of VAChT. Ct values used were the means of four to five independent repeats of 10 CNS per sample. Control gene was rp49. Primers (5' to 3') were as follows: rp49, CCAGTCGGATCGATCGATATGCTA and ACGTTGTGCACCAGGAA; VAChT, CTCATCCTCGTGATTGTA and ACGGGTATGATCTTTCC.

Statistics
LD 50 values were calculated by fitting a regression analysis to the relationship between % mortality and log[dose] using the following equation: Response ¼ 100 1þ10 ðLogLD 50 À XÞ . Statistical significance between group means was assessed using either a Student's t-test (where a single experimental group is compared to a single control group), a one-way ANOVA followed by Bonferroni's post-hoc test (multiple experimental groups) or a Chi-Square test. In all tests, confidence intervals of Ã P 0.05, ÃÃ P 0.01, and ÃÃÃ P 0.001 and ÃÃÃÃ P<0.0001 were used for significance. Data shown is mean ± s.e.m.

Increased VAChT or the presence of VAChT Y49N infers resistance to 5Cl-CASPP
Transgenic up-regulation (using GAL4/UAS) of either VAChT or VAChT Y49N , in a wildtype background, is sufficient to provide resistance to 5-Cl-CASPP [25]. To validate the previously published observation of 5-Cl-CASPP resistance when VAChT is up-regulated, and to ascertain resistance associated with endogenously expressed VAChT Y49N , mortality assays were conducted in larvae where VAChT was up-regulated in all cholinergic neurons (cha B19 >VAChT) or expression of VAChT Y49N achieved through CRISPR knock-in (VAChT Y49N/Y49N and VAChT Y49N/+ ).

Elevated VAChT increases frequency of spontaneous release events
To investigate the consequence for cholinergic synaptic function associated with observed 5-Cl-CASPP resistance, we undertook patch-clamp recordings from well-characterized aCC/ RP2 motoneurons. These neurons receive identical cholinergic synaptic input [41]. We recorded spontaneous minis, achieved by blocking action potential-dependent activity with TTX. We have previously shown that exposure to 5-Cl-CASPP significantly reduces mini frequency, but does not affect mini amplitude [30]. In order to validate these data, and to provide a baseline for comparison with the experiments conducted as part of this study, we repeated this analysis. We observed the same results. 5-Cl-CASPP (10 μg/ml) mediated inhibition of VAChT results in a significant reduction in mini frequency (38.8 ± 5.6 vs. 13.2 ± 2.9 per min, vehicle vs. 5-Cl-CASPP respectively, P = 2 x 10 −3 , Fig 2A and 2B) but no effect to mini amplitude (P = 0.91, Fig 2A and 2B). By contrast, up-regulation of wildtype VAChT is sufficient to increase mini frequency, without change to amplitude [30]. Exposure of this genotype (cha B19 >VAChT) to 5-Cl-CASPP also results in a significant reduction in mini frequency (86.8 ± 19.3 vs. 31.0 ± 6.3 per min, vehicle vs. 5-Cl-CASPP respectively, P = 0.01) and no change to amplitude (P = 0.94, Fig 2A and 2B). However, the reduction in absolute level of mini frequency, observed following exposure to CASPP, is to a degree that is not significantly different to untreated CS (treated with vehicle, P = 0.99).
Exposure to 5-Cl-CASPP also dramatically reduces the frequency of spontaneous rhythmic currents (SRCs), representative of action potential-dependant mass vesicle exocytosis (19.2 ± 4.1 vs. 3.1 ± 1.1 per min, vehicle vs. 5-Cl-CASPP respectively, P = 0.03) with no effect to SRC amplitude (P = 0.58, Fig 2C and 2D). This is suggestive of an inability of the cholinergic premotor interneurons to maintain wildtype evoked SVs release in the presence of 5-Cl-CASPP. The lack of effect to SRC amplitude, moreover, suggests that the premotor terminals release a relatively fixed number of SVs per release event: if the required number of 'full' SVs is not available, no evoked release seemingly occurs. The reduction in SRC frequency is rescued by the up-regulation of wildtype VAChT (24.1 ± 3.8 vs. 23.5 ± 4.9, vehicle vs. 5-Cl-CASPP respectively, P = 1), again with no obvious effect on SRC amplitude (P = 1). Taken together, these data suggest that upregulation of VAChT is sufficient to infer CASPP-resistance through maintenance of cholinergic SV release above a threshold level (approximately equal to that observed in CS/CRISPR control lines).  Fig 2. VAChT level modulates frequency of spontaneous release events. (A) Representative traces of minis recorded from L3 aCC/RP2 in either CS or cha B19 >VAChT (fed acetone or 10 μg/ml 5-Cl-CASPP). Scale Bars (full trace: 10pA/ 20sec, single mini: 5pA/5ms) (B) 5-Cl-CASPP-dependent down-regulation of VAChT activity in CS (grey bars) results in a reduction in mini frequency (38.8 ± 5.6 vs. 13.2 ± 2.9 per minute, vehicle vs. 5-Cl-CASPP respectively, P = 2 x 10 −3 ) with no effect to amplitude (P = 0.91). Exposure of a resistant background (blue bars, achieved by up-regulation of VAChT: cha B19 >VAChT) to 5-Cl-CASPP also causes a significant reduction in mini frequency (86.8 ± 19.3 vs. 31.0 ± 6.3 per minute, vehicle vs. 5-Cl-CASPP respectively, P = 0.01). However, the reduction in absolute level of mini frequency is to a level not significantly different to CS (treated with vehicle). Exposure of the resistant background to 5-Cl-CASPP had no effect on mini amplitude (P = 0.94) (C) Representative traces of SRCs recorded from L3 aCC/RP2 in GAL4/+ and UAS/+ parental stocks and cha B19 >VAChT (fed acetone or 10 μg/ml 5-Cl-CASPP). Scale Bar (400pA/ 20sec). (D) 5-Cl-CASPP dramatically reduces SRC frequency in parental controls (19.2 ± 4.1 vs. 3.1 ± 1.1 per min, VAChT Y49N/Y49N , respectively, P = 0.02 and 0.04, Fig 3A and 3B). By contrast, no effect on mini amplitude was observed, either for heterozygous or homozygous VAChT Y49N (Fig 3B). The observed resistance to 5-Cl-CASPP induced by expression of VAChT Y49N may thus be due to the maintenance of cholinergic mini release frequency above a critical threshold value (as described above). To test this we measured mini frequency and amplitude in the presence of 5-Cl-CASPP. As expected, mini frequency was significantly reduced in controls following ingestion of 5-Cl-CASPP (10 μg/ml, 38.8 ± 4.2 vs. 14.3 ± 1.8 per min, P = 1 x 10 −4 , Fig 3C and 3D). Mini frequency was also significantly reduced in the heterozygous VAChT Y49N/+ following exposure to 5-Cl-CASPP (65.2 ± 10.8 vs. 30.3 ± 4.8 per min, vehicle vs. 5-Cl-CASPP respectively, P = 0.01), but to a level that was not significantly different to controls fed vehicle (38.8 ± 4.2 vs. 30.3 ± 4.8 per min, P = 1). Homozygous expression of VAChT Y49N/Y49N is, however, sufficient to completely prevent 5-Cl-CASPP from reducing mini frequency (69.3 ± 20.5 vs. 61.0 ± 18.1 per min, vehicle vs. 5-Cl-CASPP respectively, P = 0.78). No significant change in mini amplitude was observed for any treatment (Fig 3E). These data are consistent with VAChT Y49N providing resistance to 5-Cl-CASPP through maintenance of a threshold frequency of cholinergic mini release. Moreover, resistance would appear proportional to mini release frequency.

Expression of Sytx 3-69 or cpx SH1 fails to provide resistance to 5-Cl-CASPP
To validate the hypothesis that spontaneous release frequency dictates resistance to 5Cl-CASPP, we identified two additional genetic approaches reported to elevate glutamatergic minis at the Drosophila NMJ. Larval mortality assays and mini recordings from aCC/RP2 were conducted following expression of sytx  or in a complexin null mutant, cpx SH1 [37, 38].

VAChT Y49N disrupts rhythmicity of cholinergic circuits
To assess how VAChT Y49N influenced motoneuron activity, we examined endogenous bursting activity (as measured by loose patch) of aCC/RP2 motoneurons. The advantage of loose patch vehicle vs. 5-Cl-CASPP respectively, P = 0.03) but with no effect to amplitude (P = 0.58). The effect of 5-Cl-CASPP on SRC frequency is rescued following up-regulation of VAChT (24.1 ± 3.8 vs. 23.5 ± 4.9, vehicle vs. 5-Cl-CASPP respectively, P = 1) with no obvious effect to amplitude (P = 1). All data points are mean ± sem, n stated in each bar.
https://doi.org/10.1371/journal.pone.0203852.g002  VAChT Y49N perturbs synaptic transmission is that the interior of the cell is not affected (the membrane remains intact). We measured both frequency of bursts and the number of individual action potentials (APs) per burst. Previous work has shown that a single SRC gives rise to a sustained depolarization (burst) that elicits multiple APs in motoneurons (22). Thus, burst frequency represents the evoked release of ACh from premotor interneurons, whilst APs fired is a postsynaptic property of the motoneurons. CS larvae exhibited 27.9 ± 7.5 bursts per minute, with an average of 15.0 ± 1.8 APs per burst. By contrast, VAChT Y49N/Y49N showed significantly reduced burst frequency (9.0 ± 4.2 per min, P = 0.04). This is predictive of a decreased ability of the premotor interneurons to maintain a normal synaptic output. However, APs per burst remained constant (12.8 ± 3.0 spikes per burst, P = 0.54) indicative that the postsynaptic motoneuron is not affected. Thus, despite an increase in spontaneous mini release (cf. Fig 3B), these data show that the VACh-T Y49N point mutation does not support 'normal' presynaptic cholinergic input to motoneurons which, in turn, disrupts motoneuron burst frequency. However, the number of APs fired by motoneurons per burst is unaffected.
We next asked whether changes to evoked cholinergic synaptic release are observable in other cholinergic neural circuits. We used the adult giant fibre system (GFS). The physiological output from stimulation of the adult GF is muscular contraction of both the tergotrochanteral muscle (TTM) and the dorsal longitudinal flight muscle (DLM) [43]. Although both DLMs and TTMs are directly innervated by glutamatergic synapses, TTM motoneurons are excited via both cholinergic and electrical synapses, whereas DLM motoneurons are excited solely be cholinergic synapses. We find that, during high frequency stimulation (100Hz), the DLM of VAChT Y49N is less able to follow a train of 10 stimuli (8.3 ± 0.7 vs. 5.8 ± 0.9, CS vs. VAChT Y49N respectively, P = 0.03, Fig 5C and 5D). The TTM, by contrast, which has additional electrical synapses, remains unaffected (9.2 ± 0.3 vs. 9.4 ± 0.4 CS vs. VAChT Y49N respectively, P = 0.60). These data further support the hypothesis that the VAChT Y49N mutation disrupts the ability of cholinergic neurons to maintain sustained evoked release.

The VAChT Y49N resistance mutation disrupts evoked vesicle release
A reduction in burst frequency of aCC/RP2 in homozygous VAChT Y49N is indicative of an inability to maintain a normal presynaptic cholinergic release. In order to address this question, we tested the ability of motoneurons to follow prolonged stimulation of premotor cholinergic interneurons. We used the blue light-sensitive ChR variant, UAS-ChR ChETA , expressed in all cholinergic neurons (ChAT BAC GAL4), to allow us to stimulate presynaptic activity. Long term stimulation at 1Hz in control lines (CS + CRISPR control) or VAChT Y49N/+ resulted in a gradual decline in evoked EPSC amplitude (these EPSPs are equivalent to SRC's but are evoked by optogenetic stimulation). Full failure of transmission occurred at 25.1 ± 1.6 vs. 28.7 ± 2.9 min (control vs. VAChT Y49N/+ , P = 1, Fig 6A and 6B). By contrast, EPSC amplitude falls off dramatically after the first stimulation in homozygous VAChT Y49N (failure of transmission occurring at 2.9 ± 1.3 min, P = 3 x 10 −4 ). A reduction in successive EPSC amplitude is clearly observable between the 1 st and 2 nd induced EPSCs, which is significantly greater in VACh-T Y49N/Y49N (90.4 ± 6.5 vs. 52.0 ± 10.6%, control vs. VAChT Y49N/Y49N respectively, P = 2 x 10 −3 ). Thus, although the amplitude of the first EPSC is not different, the ability to support a full-size second EPSC is significantly compromised in VAChT Y49N . Fig 6C displays ratio (%) of EPSC amplitude difference (2/1) at time intervals 0.1, 1, 2 and 8 seconds following initial stimulation. Regression analysis suggests the minimum time interval required between successive stimulations to generate a second full size EPSC in VAChT Y49N/Y49N is approximately doubled compared with control lines (2.4 vs. 4.8 s, control vs. VAChT Y49N/Y49N respectively). This is entirely consistent with the reduced frequency of burst activity recorded in these same motoneurons by loose-patch (c.f. Fig 5A and 5B).

Discussion
We report neurophysiological consequences resulting from two potential mechanisms of CASPP-resistance. Resistance obtained through VAChT up-regulation correlates to an increase in the availability and/or probability of SV release and not through alteration of the amount of neurotransmitter loaded per vesicle (i.e. quantal size). The lower bounds of sustainable activity appear to be dictated by a minimum threshold for release frequency, below which normal cholinergic release is unsustainable and the likelihood of mortality increases. However, resistance associated through the second route, endogenous expression of the point mutation VAChT Y49N , although showing a similar spontaneous release profile, further uncovers significant deficiencies in evoked neurotransmitter release. We attribute this to an inability of the point mutation to maintain sustained evoked SV release, perhaps indicative of a fitness tradeoff between protein functionality and insecticide resistance.
Insecticide resistance alleles commonly produce negative effects on multiple life-traits (including larval development, sex ratio, oviposition/fecundity, mass, embryonic viability and adult longevity) [44][45][46] and in some cases these are proportional to resistance level [46]. However, exceptions have been reported suggesting interactions between resistance modality and target protein functionality dictates complex life-traits [47]. We show in this study that although resistance may be attained through VAChT up-regulation, adult longevity is reduced. Reduced lifespan is also associated with transgenic up-regulation of VGLUT, where adult male longevity is reduced by >50% [18]. VGLUT mediated excitotoxicity and neurodegeneration may result from increased quantal size (i.e. excess glutamate release) [18,48]. Although, similar post-synaptic degeneration may occur in central cholinergic neurons following up-regulation of VAChT, the degree of excitotoxicity is predicted to be lower due to the inability of this transporter to affect quantal size (only affecting release frequency) [30]. Cash et al show that, in aCC motoneurons, either VAChT up-regulation or pharmacological inhibition of VAChT through 5-Cl-CASPP treatment is not sufficient to alter the postsynaptic response to perfused ACh [30]. However, up-regulation of the nAChR D7 subunit does show an increase in postsynaptic response to perfused ACh [30], suggesting that postsynaptic receptor kinetics can be modulated at central cholinergic synapses, however, not through genetic or pharmacological manipulation of pre-synaptic VAChT activity. Given the similarities in spontaneous release kinetics between VAChT up-regulation and the VAChT Y49N mutation, together with the observation that the Y49N mutation reduces endogenous and evoked activity, we rationalize it unlikely that the Y49N point mutation or VAChT up-regulation would cause a consequential alteration in postsynaptic kinetics or precipitate acute cholinergic neurotoxicity. Following a similar logic, the reduced evoked release of SVs (i.e. SRCs) in VAChT Y49N may underlie the increased longevity observed in this genotype; reduced transmission being protective. The inability to influence quantal size, at central cholinergic synapses, through up-regulation of VAChT is indicative that vesicle-loading obeys the proposed "set-point" SV loading model [30,49]. This model dictates that the level to which SVs are filled is pre-set and, thus, cannot be influenced through an increased loading rate (which might be expected following increased transporter expression). Several other studies are in agreement with this conclusion [30,50]. However, studies at mouse, rat and nematode NMJs link modulation of VAChT activity with corresponding changes in quantal size; conforming to the steady-state SV filling model [14,[51][52][53]. Furthermore, increased VGLUT at glutamatergic synapses at the Drosophila NMJ is sufficient to increase quantal size through the manipulation of vesicle size [48], an effect also not observed following up-regulation of VAChT expression [30]. Collectively, these studies are indicative of a significant difference between either SV loading at central vs. peripheral synapses and/or differences between cholinergic and glutamatergic synapses.
Our findings in this study are indicative of a direct relationship between SV release probability and pharmacological and/or genetic regulation of VAChT. Our data suggest, moreover, that block of VAChT through CASPP, results in a reduction in SV release frequency. This finding agrees with previously published data [30] and correlates with CASPP lethality [25]. Upregulation of VAChT or expression of VAChT Y49N mediates resistance to CASPP, an effect that is associated with the maintenance of spontaneous SV release above a critical threshold. The resistance effect of up-regulation of wildtype VAChT could be simply explained by an increase in the number of vesicles that escape block of VAChT by CASPP and are thus able to reach setpoint. The resistance effect of replacing the wildtype VAChT with the Y49N mutant form however, might be explained by a combination of different factors: Expression of VAChT Y49N may confer resistance to CASPP because binding of the insecticide to the transporter is prevented, because binding of CASPP to VAChT Y49N no longer prevents ACh transport and therefore does not critically reduce the number of vesicles reaching set point, and/or because binding and inhibition of VAChT Y49N function by CASPP is compensated by an increased release probability conferred by the Y49N mutation. Further experiments would be required to confirm or rule out the individual contributions of these factors.
The expression of VAChT Y49N additionally disrupts the ability of the cholinergic system to maintain sustained evoked release. This is evidenced by a reduction in frequency of endogenous bursts of AP firing in motoneurons and by a failure of the DLM to follow high frequency stimulation. The mechanism underlying this may be heightened spontaneous mini release, which in turn reduces the pool of SVs available for evoked release. This conclusion is in agreement with work at the glutamatergic Drosophila NMJ which reports complexin null mutants display a reduction in EPSPs (equivalent of SRCs) together with increased mini frequency [38]. However, the syntaxin mutant (syx  ) and a SNAP-25 mutant (SNAP25 ts ) maintain higher mini frequency together with EPSPs that are increased in amplitude [37,54].
CRISPR-dependent expression of VAChT Y49N is sufficient to increase frequency of minis and, also, to prevent a normal pattern of evoked release of ACh from premotor cholinergic interneurons. Only the former results following the up-regulation of wildtype VAChT level. In addition to suggesting that the contribution of this transporter to spontaneous and evoked release may be separable, it identifies a region of the transporter that is seemingly required for sustained evoked release. The VAChT Y49N mutation disrupts a region resembling the wellcharacterized trafficking motif YXXØ (with Ø being any large hydrophobic amino acid): specifically causing a YMVI to NMVI mutation. This motif, amongst others (e.g. YNYY in mammals) is known to be involved in protein sorting, the Y residue being highlighted as the most critical determinant [55]. The positioning of this signal on a cytosolic region of VAChT indicates that this region could potentially be involved in trafficking. Disruption of the YNYY motif in mammalian VAChT is sufficient to cause mis-localisation and retention of the transporter to the plasma membrane [56,57]. Whilst mis-localisation or membrane retention of VAChT may be sufficient to perturb sustained evoked release, it does not offer a plausible explanation why spontaneous vesicle release is increased in these mutants and following upregulation of wildtype VAChT. It is notable in this regard that a glycine-to-arginine substitution at position 347 (G347R) disrupts a reported VAChT-synaptobrevin interaction in C. elegans that impacts SV release probability [58]. Glycine is conserved in Drosophila VAChT at position 342 suggesting a potential similar involvement in the SNARE assemblage.
Up-regulation of genes relating to insecticide detoxification (e.g. glutathione S-transferase and cytochromes P450s) is a common form of insecticide resistance [59][60][61][62]. Natural expressional modification of target proteins (expression level or mutation) are rarer but have been found to confer resistance in other cholinergic targets. For example, upregulation of acetylcholinesterase (ace-1) is reported in organophosphate and pyrethroid resistant field populations in the spider mite (Tetranychus evansi) [63] in addition to gene duplication of ace-1 in mosquito (Culex pipiens) [64]. This study shows that expressional modification of a target protein (VAChT) or the acquisition of a point-mutation (Y49N) is sufficient to confer resistance to CASPP. This mutation affects a characterized motif in VAChT (YMVI) that is seemingly critical for high frequency SV evoked release. Further characterisation of this motif may provide a better understanding of how this transporter contributes to evoked synaptic release, but may also identify a novel target for insecticidal attack.