Indole inhibits AaegOR8 activation by (R)-1-octen-3-ol.
Based on the chemical similarities between indole and DEET (Fig. 1A), we surmised that indole exerts an inhibitory effect on AaegOR8. A 10-3 M indole concentration reduced by approximately 30% the AaegOR8 current amplitude elicited by 10-7 M (R)-1-octen-3-ol (Fig. 1B). This effect was not observed in Orco-injected oocytes.
To determine the nature of this inhibitory effect, we established a series of concentration-response curves using OR8-Orco-injected oocytes exposed to increasing concentrations of (R)-1-octen-3-ol alone or combined with 10-4 M, 10-3 M or 10-2 M indole (Fig. 1C). Two types of currents were observed, including the expected agonist-induced depolarization currents and unusual indole-dependent hyperpolarization or reduction in baseline currents.
The interpolated EC50 values for octenol alone versus 10-4 M and 10-3 M indole were not statistically significant (Fig. 1D, Additional file 5: Table S2). The EC50 value elicited by 10-2 M indole was significantly different from (R)-1-octen-3-ol alone but was moderate. These findings suggest that indole does not have an important effect on the sensitivity of this receptor for (R)-1-octen-3-ol. Contrary to our initial results (Fig. 1B), we did not observe any significant inhibition of the amplitude response as the indole concentration increased (Fig. 1E). However, current amplitudes are contingent on oocyte inherent variability. To address this limitation, we determined the systematic effect of indole on depolarization current amplitudes by normalizing all the current responses to the initial 10-7 M (R)-1-octen-3-ol exposure (Fig. 1C). 10-2 M indole consistently reduced the current amplitude across the concentration of (R)-1-octen-3-ol, excluding 10-10 M. We confirmed that 10-3 M indole significantly reduced the response amplitude of AaegOR8 to 10-7 M (R)-1-octen-3-ol (Fig. 1F).
Indole concentration of 10-3 M and 10-2 M elicited hyperpolarization currents (reductions in current baseline) in the presence of (R)-1-octen-3-ol concentrations ranging from 10-10 M to 10-7 M (Fig. 1C). These hyperpolarization currents were concentration-dependent and were surmounted at higher (R)-1-octen-3-ol concentrations (Fig. 1G). To understand the contribution of indole alone to these currents, we exposed OR8 to increasing concentrations of indole (Fig. 2A). 10-6 to 10-4 M indole evoked small depolarization currents. The two highest indole concentrations elicited either depolarization or hyperpolarization currents that were a fraction of the initial 10-7 M (R)-1-octen-3-ol stimulation. To better characterize these small yet inconsistent effects, we focused on the currents elicited by 10-2 M indole and consistently observed these small hyperpolarization and hyperpolarization currents (Fig. 2B). By comparison, this same concentration of indole, in the presence of (R)-1-octen-3-ol at concentrations as low as 10-10 M, elicited currents reaching the initial response to 10-7 M (R)-1-octen-3-ol, suggesting that the significant indole-induced hyperpolarization currents require the presence of (R)-1-octen-3-ol (Fig. 2C).
Indole modifies the OR8-mediated current baseline of the oocyte membrane.
The cause for the observed hyperpolarization currents caused by indole in the presence of (R)-1-octen-3-ol was intriguing. It mirrored a phenomenon previously documented with AaegOR8, AaegOR2 and AaegOR10 [4,6] and more recently with additional ORs from Culex quinquefasciatus, Aedes aegypti, and Anopheles gambiae [24]. To investigate whether the hyperpolarization current was a transient response or a durable modification of the current baseline, we exposed OR8-injected oocytes to a change of perfusion buffer by switching from the ND96 solution to a 5.10-3 M indole perfusion. We also administered increasing ten-fold dilutions of (R)-1-octen-3-ol before reversing the perfusion solution back to ND96 (Fig. 3A). Prior and after the two perfusion buffer exchanges, the oocyte was exposed with a transient stimulation of 10-7 M (R)-1-octen-3-ol for control purposes. The switch from ND96 to indole elicited a stable decrease in the baseline current not observed in water-injected oocytes (Fig. 3A). (R)-1-Octen-3-ol produced very little depolarization currents at all tested concentrations. By
comparison, DEET and IR3535 evoked larger currents. The opposite buffer switch exhibited a stable decrease in baseline current most pronounced in the case of DEET and IR3535 as well (Fig. 3A). In these experiments, we treated the oocytes with a lower indole concentration, as compared to DEET and IR3535, because 10-2 M indole consistently killed the perfused oocytes. All observed currents elicited by (R)-1-Octen-3-ol, indole, a mixture of these two ligands, and buffer switch are summarized in Fig. 3B.
Indole inhibits close-range human-host attraction.
To explore the behavioral role of indole, we exposed a human hand to female mosquitoes using an arm-in-a-cage assay (Fig. 4A insert). The hand was covered with a protective glove allowing mosquitoes to detect human skin odor through a window created by an open area on the dorsal side of the hand (Additional file 6: Video S1). This open area was protected by a screen and was equipped with an odor delivery system (Additional file 1: Fig. S1., Additional files 2&3). Increasing doses of indole ranging from 10-6 to 10-1 M were deposited on this delivery system and repellency was measured in terms of number of mosquito visits and duration of visits. The repellency effect of DEET was significantly different from vehicle and indole 10-6 M (Fig. 4A) (Kruskal-Wallis H test, H = 106.11, df = 8, p-value < 0.0001). All indole treatments, except 10-6 M, were significantly different from the vehicle (Fig. 4A). Increasing indole doses reduced the number of mosquito visits from 40.6 to 93.8%. We observed a 3.6% inhibition with a 10-6 M indole concentration but this effect was not statistically significant. Looking at the accumulated landing numbers, vehicle and 10-6 M indole elicited overlapping temporal dynamic (Fig. 4B). Indole at 10-1 M had a significantly higher temporal repellency than all other treatments including DEET at the same concentration. Other indole treatments exhibited intermediate temporal repellency between these two extremes. However, the only significant differences were observed between the vehicle and 1M indole (Additional file 7: Fig. S2). In terms of visit durations, mosquitoes spent on average the same amount of time on the open area when landing occurred (Fig. 4C) (ANOVA, F(8,18) = 1.219, P = 0.343).
Indole reduces 1-octen-3-ol-mediated attraction.
While we have circumstantial pharmacological evidence that indole may in part affect OR8-mediated detection, we do not have any direct indication that indole affects 1-octen-3-ol-mediated attraction in the context of human host-seeking. To explore this possibility, we used a flight tunnel (Fig. 5A) to expose human-host seeking female mosquitoes to a synthetic blend composed of CO2 and 1-octen-3-ol. We used three odor treatments, including CO2 alone or in combination with 1-octen-3-ol and indole. We divided the ROI into three sections (ROI-1,2,3, Fig. 5A) to explore possible differences in terms of trajectory speed, velocity and tortuosity. Kernel density estimations of mosquito locations along the X-axis were statistically different between the three treatments (Fig. 5B). A bird eye view (X-Y plane) of flight trajectories representing flight speed suggested differences between the treatments (Fig. 5C, see example of individual trajectories in Additional file 8: Fig. S3). 1-Octen-3-ol seemed to increase the number of trajectories and coverage of the ROI while indole appeared to reduce flight speed across that same area. Among all three ROIs, speed was higher and more consistent in response to CO2 in ROI-1-2 than in ROI-3 (Fig. 5D) (Kruskal-Wallis H test, H = 86.8, df = 2, P < 0.0001). As a result, we focused on ROI-1&2 for further analyses. As reflected in Fig. 5C, the addition of 1-octen-3-ol elicited higher speeds than with CO2 alone (Fig. 5E). The addition of indole elicited significant decreases in speed compared to those observed with CO2 alone or in combination with 1-octen-3-ol (Kruskal-Wallis H test, H = 301.2, df = 2, P < 0.0001). While upwind velocity did not show any statistical differences between CO2 and 1-octen-3-ol, indole elicited lower upwind velocities (Fig. 5F) (Kruskal-Wallis H test, H = 35.5 , df = 2, P < 0.0001). Finally, indole-induced tortuosity was significantly higher than those elicited by CO2 alone or in combination with 1-octen-3-ol (Fig. 5G) (Kruskal-Wallis H test, H = 20.9 , df = 2, P < 0.0001).