Dibenzo[b,e]oxepin-11(6H)-one (named as doxepinone) was synthetized following the protocol developed by our group (Scoccia et al., 2017 and Supplementary Figure S1). Briefly, 2-(phenoxymethyl) benzoic acid was prepared by treating the commercially available isobenzofuran-1(3H)-one with sodium phenoxide, which was obtained by reacting phenol with NaH in the presence of DMF at reflux. 2-(phenoxymethyl) benzoic acid was then cyclized by intramolecular acylation from the carboxylic acid compound by using FeCl₂ and dichloromethyl methyl ether as cooperative system in the presence of dichloromethane at room temperature. Purity was determined by elemental analysis and melting point.

L-Glutamic acid monosodium salt hydrate, ivermectin and dimethyl sulfoxide were from Sigma-Aldrich Chem Co.

Nematode strains used were: N2: Bristol wild type; and the null mutants of Cys-loop receptor subunits: RB918: acr-16(ok789); DA1316: avr-14(ad1302);avr-15(ad1051);glc-1(pk54); CB382: unc-49(e382); CB904: unc-38(e264); MT9668: mod-1(ok103). All strains were obtained from the Caenorhabditis Genetic Center, supported by the National Institutes of Health - Office of Research Infrastructure Programs (P40 OD010440). Nematodes were maintained at 21°C using standard culture methods (Brenner, 1974; Stiernagle, 2006; Hernando et al., 2012, 2019). Assays were carried out following standard protocols described in WormBook¹.

All behavioral assays were done at room temperature (21–23°C) and all comparisons were done in parallel. Assays were performed with young adult hermaphrodite worms from synchronized plates. For comparison among different drug concentrations or strains, the assays of the control and different conditions were performed simultaneously. For each condition, 30 worms (n = 30) were used in paralysis assays and 20 worms (n = 20) in thrashes assays. Each condition was evaluated 4 times in different days with different worm batches and always in parallel with the control, as described before (Hernando et al., 2012, 2019).

Thrashing assays were performed in 100 μl M9 buffer in the absence or presence of the drug in a 96-well microliter plate as described before (Jones et al., 2011). A single thrash was defined as a change in the direction of bending at the mid body. All assays were carried out by two independent operators and were blinded to the sample identities.

Paralysis was determined on agar plates containing the tested drug at room temperature as described before (Hernando et al., 2012). Body paralysis was followed by visual inspection at the indicated time (up to 120 min) and was defined as the lack of body movement in response to prodding.

For prodding, we used the gentle touch stimulus delivered to the body with an eyebrow hair, avoiding touching the animals too near the tip of the head or tail (Hernando et al., 2019). We evaluated different types of nematode paralysis: flaccid paralysis in which worms appear lengthened; spastic paralysis in which worms appear shorter, and stationary paralysis in which worms are immobile but respond to prodding by contracting body wall muscle (Kass et al., 1980; Hernando et al., 2019). Videos were acquired with a digital camera ToupCam UCMOS 05100KPA (Toup Tek Photonics).

Stock solutions of ivermectin (IVM) and doxepinone were in dimethyl sulfoxide (DMSO). For all assays, the final concentration of DMSO was lower than 1%.

Measurements were performed with young adult worms on agar plates. Plates contained 1 μM IVM, 50 or 100 μM doxepinone. DMSO was used as the vehicle and was present in all control plates at concentrations lower than 1% v/v. Wild-type and DA1316 young adult worms were transferred to drug plates and allowed to remain at 20°C for a 30 min period. We performed 3 independent whole experiments with different worm batches. Each experiment included comparison of the effects of drugs on wild-type and mutant strains in parallel, with n = 14 animals for each condition in each experiment. The number of contractions in the terminal bulb of the pharynx (pumps per minute) was counted using a stereomicroscope at 50× magnification.

Young adult hermaphrodite worms (n = 10 worms for each condition) were transferred to NGM plates containing 2.5 mM doxepinone. Vehicle (1% DMSO) was used as a control. After 2 h, images were acquired with a digital camera ToupCam UCMOS 05100KPA (Toup Tek Photonics) and body length was measured using FIJI-ImageJ software. Four independent whole experiments were analyzed in parallel with the controls.

GluCls were transiently expressed in BOSC 23 cells, which are modified HEK 293T cells (Pear et al., 1993). The cDNAs encoding the C. elegans GluClα1 (containing gfp between transmembrane domains M3 and M4) and GluClβ subunits, both subcloned into the pcDNA3.1 vector, were kindly provided by Dr. Paas (Degani-Katzav et al., 2016). Cells were transfected by calcium phosphate precipitation with the subunit cDNAs (total 4 μg/35 mm dish) at a ratio GluClα1:GluClβ 1:1 essentially as described before (Bouzat et al., 2008; Nielsen et al., 2018). All transfections were carried out for about 8–12 h in DMEM with 10% fetal bovine serum and were terminated by exchanging the medium. Cells were used for whole-cell recordings 2 or 3 days after transfection, time at which maximal functional expression levels are typically achieved (Bouzat et al., 2008).

Macroscopic currents were recorded in the whole-cell configuration as described previously (Bouzat et al., 2008; Corradi et al., 2009).

The pipette was filled with intracellular solution (ICS) containing 134 mM KCl, 5 mM EGTA,1 mM MgCl₂, and 10 mM HEPES (pH 7.3). The extracellular solution (ECS) contained 150 mM NaCl, 0.5 mM CaCl₂, and 10 mM HEPES (pH 7.4). After the whole cell formation, ECS containing the agonist or drug was rapidly applied using a three-tube perfusion system with elevated solution reservoirs for gravity-driven flow and switching valves controlled by a VC3 controller (ALA Scientific). The solution exchange time was estimated by the open pipette method as described by Liu and Dilger (1991). This method consists in applying a pulse of 50% diluted ECS to an open patch pipette, which produces a sudden change in the current measured by patch-clamp amplifier. After proper adjustment of the electrode position, the current jump in our system varied between 0.1 and 1 ms (Corradi et al., 2009; Andersen et al., 2016). The compound doxepinone was dissolved in ECS from DMSO stock solutions. The final concentration of DMSO used to solubilize doxepinone was lower than 0.2%. To study the modulatory action of doxepinone, responses were evaluated following co-application or preincubation protocols. After whole cell formation, 3 mM glutamate-elicited currents (control currents) were first recorded by a pulse (6 s) of ECS containing glutamate. The compound was then co-applied with glutamate (Co-application protocol) or applied during 1 min in the absence of glutamate before the second glutamate pulse (Preincubation protocol). For all experiments, the duration of the glutamate pulse was 6 s and the time of recording was 8 s. A 60-s wash with ECS alone allowed total recovery of control currents. The treated currents were normalized to currents elicited by glutamate alone in the same cell (control current). At the end of the protocol, the control current was again tested and the experiments in which the currents were reduced to more than 80% of the original control current were discarded.

Currents were filtered at 5 kHz and digitized at 20 kHz using an Axopatch 200B patch-clamp amplifier (Molecular Devices, CA, United States) and acquired using WinWCP software (Strathclyde Electrophysiology Software, University of Strathclyde, Glasgow, United Kingdom). The recordings were analyzed using the ClampFit software (Molecular Devices, CA, United States). Currents were fitted by a single exponential function according to the equation:

I(t) = I[exp(-t/τ_d)] + I_∞

in which t is time, I is the peak current, I_∞ is the steady state current value, and τ_d is the decay time constant. Net charge was calculated by current integration (Andersen et al., 2016). The rise time (tr_10–90%) corresponds to the time taken by the current to increase from 10 to 90% of its maximal value.

Experimental data are shown as mean ± SD. Statistical comparisons were done using two-tailed Student’s t-test for pairwise comparisons or oneway ANOVA followed by Bonferroni’s post hoc tests for multiple comparisons. All the tests were performed with SigmaPlot 12.0 (Systat Software, Inc.). Statistically significance was established at p-values < 0.05 (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001). Concentration–response curves were determined by non-linear regression fits to the Hill equation using Prism 5.0 (GraphPad, San Diego, CA, United States).

We have previously found that dibenzo[b,e]oxepin-11(6H)-one, referred to as doxepinone, produced rapid paralysis of C. elegans in liquid medium (Scoccia et al., 2017). After 10 min exposure, the compound produced a concentration-dependent decrease of the thrashing rate (IC₅₀ ∼300 μM) (Scoccia et al., 2017). Because doxepinone is an interesting synthetic molecule with potential anthelmintic activity we sought to explore its anthelmintic effects and identify target sites.

To determine receptor targets involved in the rapid effects of doxepinone on C. elegans swimming rate, we explored the effects on selected mutant strains lacking Cys-loop receptors involved in worm locomotion, which are targets of anthelmintic drugs. The screening is based on the hypothesis that the absence of the target site leads to drug resistance and, therefore, in the mutant worm, the thrashing rate in the presence of doxepinone will not be affected. Since some mutant strains show uncoordinated phenotypes, we determined the number of thrashes/min of wild-type and each mutant strain in liquid medium in the absence and presence of doxepinone (Figure 1A). For each strain, 20 worms in each condition were used, and experiments were repeated with different worm batches and in different days in four independent assays, always in parallel with the control condition.

The thrashing rate of wild-type worms in M9 buffer (plus 1% DMSO) was 204 ± 9.3 min^–1 (Figure 1A). After 30 min pre-exposure to 0.1 mM doxepinone, this rate decreased about 50% (103 thrashes/min, p < 0.001, Student t-Test) (Figure 1A).

Mutant worms lacking the UNC-38 subunit (CB904 strain), which is an essential L-AChR subunit, or lacking UNC-49 (GABA) receptors (CB382 strain) showed lower thrashing rates than wild-type worms as well as uncoordinated phenotypes (Brenner, 1974, Figure 1A). Nevertheless, doxepinone reduced the trashing rate in both mutants (Figure 1A, p < 0.001), with the magnitude of the reduction being similar to that observed in wild-type worms (Figure 1B). The MT9668 strain lacks the serotonin-gated chloride channel, MOD-1, that is involved in locomotion and behavior (Ranganathan et al., 2000; Komuniecki et al., 2012). This mutant strain was sensitive to the drug (Figure 1A, p < 0.001), which produced a decrease of the trashing rate similar to that observed in wild-type worms (Figure 1B). Thus, L-AChR, UNC-49 and MOD-1 receptors are not the main receptors involved in the rapid effects of doxepinone in the swimming rate. The RB918 strain, which lacks the nicotine-sensitive nAChR (N-AChR) present in muscle (Touroutine et al., 2005), was also sensitive to doxepinone (Figure 1A, n = 20 worms for each condition). However, the decrease of the trashing rate was slightly, but statistically significantly, lower than that observed for wild-type worms, indicating some type of contribution of this receptor to doxepinone action (Figure 1B, p < 0.01).

The thrashing rate of the triple mutant worms lacking three GluCl subunit genes (DA1316) was similar to that of wild-type animals in the absence of the drug (Figure 1A). Interestingly, exposure to 0.1 mM doxepinone did not affect this rate, in contrast to the effects observed in wild-type worms (Figure 1A, n = 20, p > 0.05). The percentage of the reduction of the thrashing rate due to the presence of doxepinone was statistically significantly different between wild-type and DA1316 worms (Figure 1B, p < 0.001). Altogether, our results indicate that GluCls are involved in the rapid effect of doxepinone on the swimming rate.

To determine the type of paralysis and the contribution of the different Cys-loop receptors to doxepinone effects, we performed paralysis assays on agar plates containing the drug.

We first characterized the type of paralysis exerted by doxepinone by exposing wild-type worms to 2.5 mM doxepinone for 30–120 min in agar plates. After 60 min exposure, the worms in the presence of doxepinone were immobile but respond to prodding by contracting body muscle (Supplementary Video S1). After 2 h exposure, worms did not show any response and were completely paralyzed. The length of worms in the absence of the drug was 1.16 ± 0.01 mm whereas it was 1.10 ± 0.04 mm after 1 h exposure (n = 10, p > 0.05), indicating neither spastic nor flaccid paralysis.

For wild-type worms, a clear concentration- and time-dependent paralysis was detected during the 2 h assay at a 1–3 mM doxepinone concentration range (Figure 2A). At the maximum exposure time (2 h), ∼85% adult worms were paralyzed by 3 mM doxepinone (Figure 2A, n = 30). The IC₅₀ values determined for the inhibition of moving worms in agar plates were 2.58 ± 0.01 mM and 2.10 ± 0.01 mM for 90 and 120 min exposure, respectively (Figure 2B). Paralysis assays on agar plates measured at short times usually require higher drug concentrations than those used in liquid medium probably since the drug must be absorbed from the solid phase.

To confirm the contribution of the different Cys-loop receptors to doxepinone effects, we next measured paralysis as a function of time of different mutant worms exposed to 2.5 mM doxepinone. We found that the reduction of moving worms as a function of time of exposure for MT9668, CB904, and CB382 strains did not differ from that of wild-type worms (Figure 2C). For these mutants, the fraction of moving worms was reduced to ∼0.40 after 2 h exposure on agar plates containing 2.5 mM doxepinone (Figure 2C). Although slight, there was a statistically significant difference in the effects of the drug on RB918 worms (lacking N-AChR) respect to wild-type worms after 2 h exposure (p < 0.01, n = 30). For worms lacking GluCl subunits (DA1316), only a slight reduction of the fraction of moving animals was detected after 2 h exposure to doxepinone (18%). This reduction was markedly different to that of wild-type worms and other mutants (∼60%), again indicating that GluCls are involved in doxepinone paralysis (Figure 2C).

Because GluCls are targets of IVM we compared the effects of doxepinone with those of IVM on DA1316 (lacking three GluCl genes) and wild-type strains. After 30 min exposure to 10 μM IVM in liquid medium, wild-type worms were fully paralyzed whereas mutant worms (DA1316) showed only 50% reduction of the thrashing rate (Figure 3A). In agar plates containing 300 μM IVM, wild-type animals showed a time-dependent paralysis that yielded 80% paralyzed worms after 2 h. On the contrary, the percentage of paralyzed mutant worms was smaller than 10% after 2 h (Figure 3B). After 2 h exposure, the difference in the fraction of moving worms between wild-type and mutant animals was statistically significative (p < 0.001, n = 30). We also determined that the type of paralysis of wild-type worms exposed to 300 μM IVM in agar plates during 60 min was similar to that observed for doxepinone: worms appeared immobile but respond to prodding by contracting body muscle (Hernando and Bouzat, 2014; Supplementary Video S2).

One of the hallmark effects of IVM involving GluCls is pharyngeal pumping inhibition (Dent et al., 1997, 2000). We therefore evaluated the effects of doxepinone on the pharyngeal pumping rate of wild-type and mutant worms lacking GluCl subunit genes (DA1316). In the absence of drugs, wild-type and DA1316 worms showed similar pumping rates of ∼180–200 min^–1 (Figure 4). It is important to note that despite lacking GluCls the mutant worms show normal pumping due to compensatory effects and that IVM also affects other Cys-loop receptors (Pemberton et al., 2001; Lynagh and Lynch, 2012; Trojanowski et al., 2016).

The exposure of worms to 1 μM IVM decreased 4-fold the pumping rate in wild-type worms and about 1.7-fold in DA1316 worms (n = 14, p < 0.001) (Figure 4). These results confirmed that worms lacking GluCl subunits are more resistant to the pumping rate inhibition by IVM than wild-type animals (Figure 4). In wild-type animals, increasing concentrations of doxepinone produced a monotonically decrease of the pharyngeal pumping rate, which was reduced 4-fold at 50 μM and 100% at 100 μM (Figure 4). On the contrary, only ∼1.5-fold reduction was observed in the DA1316 worms in the presence of 100 μM doxepinone with respect to the control, thus indicating that the mutants are resistant to the pharyngeal pumping effect of doxepinone (Figure 4). Thus, the actions of doxepinone correlate with those of IVM and depend on GluCls.

To confirm that doxepinone acts at GluCls and to unravel the mechanism by which it may affect these receptors, we expressed GluClα1/β in BOSC 23 cells and measured whole-cell currents at −60 mV holding potential. Currents were elicited by rapid application of 3 mM glutamate, which is a concentration higher than its EC₅₀ for these receptors (∼1.5 mM) but lower than that required for saturation (Cully et al., 1994; Degani-Katzav et al., 2016).

A 6-s pulse of 3 mM glutamate in ECS elicited macroscopic currents that reached the peak with a rise time of 49.30 ± 16.20 ms (n = 9) and decayed in the presence of the agonist due to desensitization. Typically, peak currents varied between 2000 and 7000 pA. The decay was fitted by a single component with a time constant of 2100 ± 1040 ms (n = 9). The application of another pulse of glutamate after a 20-s wash with ECS allowed full recovery of the peak current, indicating that most receptors recovered from desensitization in the absence of the agonist after this period (Figure 5A).

To further characterize GluCl-mediated currents, we constructed current-voltage relationships by measuring the peak current elicited by 3 mM glutamate as a function of the holding potential (Figure 5B). As shown in the figure, the magnitude of GluCl-elicited currents increased linearly with the voltage, indicating an ohmic behavior, and currents did not show important rectification (n = 4).

We next proceeded to decipher the molecular actions of doxepinone at GluCls. To first determine if doxepinone can activate GluCls, a 6-s pulse of 0.5 mM doxepinone (n = 7) or 1 mM doxepinone (n = 3) was applied to cells in the whole-cell configuration at -60 mV. No currents were elicited by doxepinone and sequential application of 3 mM glutamate to the same cell elicited macroscopic responses, indicating the presence of functional GluCls (Figure 5C). Thus, doxepinone does not act as an agonist of GluCls.

We next evaluated the action of doxepinone as a modulator of glutamate-activated currents. To this end, we used two different drug application protocols, one including preincubation of the drug before glutamate application (Preincubation protocol) and the other, application of doxepinone together with glutamate (Co-application protocol).

For the preincubation protocol, a pulse of ECS containing 3 mM glutamate (6 s-pulse, control current) was first applied to the cell held at -60 mV, and the cell was incubated during 1 min with ECS containing doxepinone (1 mM) before a second pulse of ECS-glutamate was applied (treated) (Figure 6A). The same protocol was repeated three times in the same cell, each time separated by a 60-s period. An illustrative example of an experiment from a single cell is shown in Figures 6A,B. The results from different cells were averaged and shown in Figure 6C.

Preincubation of the cell with 1 mM doxepinone produced a slight decrease of the peak current and a profound decrease of the current decay time constant and net charge (Figure 6C). Compared to the corresponding control current in each cell, the peak current was reduced to 0.81 ± 0.08 (p = 0.001), the decay rate to 0.57 ± 0.13 (p = 0.001), and the net charge to 0.49 ± 0.11 (p = 0.002, n = 6). Thus, the main effect of the drug on glutamate-elicited currents is the increase in the decay rate (or decrease of the decay time constant), which leads to a concomitant decrease in the net charge. We also determined that the percentage of changes after drug application were similar among the three applications on each cell. Also, no significant differences were found in the relative peak current (1.01 ± 0.10), net charge (0.99 ± 0.17), and decay time constant (0.95 ± 0.06) with respect to the control when preincubation was performed with 0.2% DMSO in ECS in the absence of doxepinone (n = 4).

We also explored if doxepinone applied together with glutamate (co-application protocol) affected GluCl currents (Figures 6E,D). In each cell, we applied three pulses of glutamate containing ECS, each one separated by 20 s, and then three pulses of ECS containing 3 mM glutamate and 1 mM doxepinone (Figure 6D). Doxepinone did not produce statistically significant changes in the peak currents (0.99 ± 0.02, p = 0.319) but produced a slight and statistically significantly decrease of the net charge (0.83 ± 0.06, p = 0.001) and decay time constant (0.76 ± 0.10, p = 0.008) (n = 5 cells) (Figure 6E). Thus, co-application of doxepinone inhibited glutamate-activated currents but the changes were smaller than those determined under the preincubation protocol.