Odorant ligands for the CO2 receptor in two Anopheles vectors of malaria

Exhaled CO2 is an important host-seeking cue for Anopheles mosquitoes, which is detected by a highly conserved heteromeric receptor consisting of three 7-transmembrane proteins Gr22, Gr23, and Gr24. The CO2 receptor neuron has been shown to also respond sensitively to a variety of odorants in Aedes aegypti. The detection of CO2 is important for upwind navigation and for enhancing the attraction to body heat as well as to skin odorants. The orthologs of the CO2 receptor proteins are present in malaria-transmitting mosquitoes like Anopheles coluzzii and Anopheles sinensis. Activators and inhibitors of the CO2-neuron were tested on the maxillary palps in these two species by single-sensillum electrophysiology. The electrophysiological testing of three prolonged-activator odorants identified originally in Aedes aegypti also showed varying ability to reduce the CO2-ellicited increase in spikes. These findings provide a foundation for comparing the functional conservation with the evolutionary conservation of an important class of odorant receptor. The identification of a suite of natural odorants that can be used to modify the CO2-detection pathway may also contribute to odor-blends that can alter the behavior of these disease transmitting mosquitoes.

Using single sensillum recordings, we screened the cpA neuron of two anopheline mosquito species, anthropophilic An. coluzzii and facultative anthropophilic Anopheles sinensis, that transmit malaria with a large set of odorants that were initially identified by chemical-informatics as putative CO 2 receptor ligands, and subsequently tested in the Ae. aegypti mosquitoes 6 . We identify several odorants that show conserved effects as activators, inhibitors, and ultra-prolonged activator of the cpA neuron. Some of these odorants have potential in reducing anopheline-biting rates.
Conservation of response to agonists in An. coluzzii. In order to test conservation of the CO 2 -receptor neuron responses to different odorants, we tested a structurally diverse set of 67 ligands of the CO 2 -receptor previously identified in Ae. aegypti and Cx. quinquefasciatus 6 . The An. coluzzii CO 2 neuron responded to several of these odorants with different chemical structures (Fig. 1B). Out of the 67 odors tested 35 (52%) evoked responses ≥30 spikes/sec whereas 14 odors (21%) showed responses lower than the solvent (paraffin oil; Fig. 1C). The odorants that evoked the strongest activation from the An. coluzzii cpA neuron were further evaluated in www.nature.com/scientificreports www.nature.com/scientificreports/ dose-response assays across a range of five orders of magnitude. All odorants showed a dose-response and four out of the five odors still evoked responses ≥30 spikes/sec with headspace from a 10 −3 dilution (Fig. 1D).

Conservation of ultra-prolonged activators in
An. coluzzii. Another class of ligands that have been identified in A. aegypti are ultraprolonged activators of the CO 2 neuron 6,13 . In order to test these longer-term responses, recordings were performed as before with three known odorants ( Fig. 2A). This analysis revealed that two of three odorants are conserved in their ability to evoke ultraprolonged activation in An. coluzzii. After a 3-s exposure to a (E)-2-methylbut-2-enal stimulus, the cpA neuron continues firing at ~50 spikes/sec for at least 5 min (Fig. 2B,C). Consistent with previous results responses to repeated CO 2 stimuli during this 5-min period after the pre-exposure to (E)-2-methylbut-2-enal were significantly reduced (Fig. 2D).

Conservation of inhibitors in An. coluzzii.
Among the odorants that induced responses lower than the solvent, six odors actually inhibited the baseline activity of the An. coluzzii cpA neuron when tested with the headspace from 10 −2 concentration odor cartridges (Fig. 1A). In order to test whether some of these odorants could constitute potential antagonists of cpA, we tested the ability of 21 odorants at a higher concentration (headspace from 10 −1 concentration odor cartridges) to inhibit CO 2 -mediated (0.15% concentration) activation of the An. coluzzii cpA neuron in overlay assays. Of the 21 odorants tested, 11 were capable of reducing CO 2 -mediated cpA activation between 20% and 45%, and 5 odorants inhibited CO 2 activation by >80% (Fig. 3A). Four top inhibitors were selected for dose-response assays. Four odorants were able to inhibit cpA activation by at least 50% when tested at 10 −2 concentration and propanal evoked similar levels of inhibition when tested at 10 −3 (Fig. 3B). Representative traces from the cp sensillum to 1 s pulses of 0.15% CO 2 prior to and following a 3-s exposure to either solvent (PO-paraffin oil) or (E)-2-methylbut-2-enal (headspace above 10% solution on filter paper). (D) Mean responses of the cpA neuron to 1 s pulses of 0.15% CO 2 , calculated by subtracting 1-s of baseline activity prior to each stimulus after exposure to paraffin oil (gray) or (E)-2-methylbut-2-enal (headspace above 10% solution on filter paper) (orange). n = 5-6 individuals; t test, ***p < 0.001. Error = s.e.m.
www.nature.com/scientificreports www.nature.com/scientificreports/ Conservation of inhibitors in An. sinensis. In order to test whether inhibitory odorants of the An.
coluzzii cpA neuron could be of utility in An. sinensis, four of the strong inhibitors (propanal, ethyl pyruvate, thiophene-2-thiol, and 4-methyl piperidine) were tested at two concentrations using electrophysiology. All the tested odorants showed some degree of inhibition, but to varying extent. The strongest inhibition was observed for thio-2-thiol and ethyl pyruvate at the higher concentrations (1%) (Fig. 4A). Among the three tested amines, amyl amine (AA) exhibited the highest inhibitory activity followed by butyl amine (BA). Spermidine (SP), on the other hand, acted as a weaker inhibitor in the presence of CO 2 (Fig. 4B). Taken together these results indicate that the An. sinensis CO 2 receptors respond similarly to Ae. aegypti and An. coluzzi when it comes to CO 2 response inhibition.

Discussion
Unlike the complex blend of human skin odor, the CO 2 in exhaled breath provides a simpler cue to study. Most mosquito species are strongly attracted to CO 2 exhaled from human breath 2 . Carbon-dioxide is detected by the cpA neuron upon binding to a receptor, comprised of three members of the gustatory receptor gene family and named GR22, GR23, and GR24 in An. coluzzii 4,5 . The detection of CO 2 plays several important roles in host-seeking behaviors like long-range navigation towards a live animal 2,14 . The identification of volatile ligands of the CO 2 receptor neuron can contribute to the design of masking agents using inhibitors and trapping-lures using activators, both of which can reduce human contact and prevent disease transmission by mosquitoes 6 . We previously developed a computational approach 15 which we applied to identify novel odorant ligands of the cpA neuron in A. aegypti 6 . Here we tested the conservation of the ligand responses in two species of anopheline mosquitoes, An. coluzzii and An. sinensis, that transmit malaria in Africa and Asia respectively. An. sinensis is one of the vectors of malaria in Asia and has Gr receptor orthologs that are closely related to ones in An. coluzzi. A high degree of functional conservation was observed amongst the ligands and we identified all three classes of ligands: activators, prolonged activators, and inhibitors. www.nature.com/scientificreports www.nature.com/scientificreports/ In general, the responses to activators were weaker in An. coluzzii than in Ae. aegypti 6 , and only two odorants ((E)-pent-2-enal and methyl acetate) evoked stronger responses in An. coluzzii than in Ae. aegypti 6 . Conversely, inhibition of CO 2 -mediated cpA activation is stronger in An. coluzzii, as odorants unable to inhibit the cpA neuron in Ae. aegypti 6 were capable of inhibiting the An. coluzzii and An. sinensis CO 2 neuron counterparts. Amongst the inhibitors we have identified both high and low volatility compounds that act in both An. coluzzii and An. sinensis.
The detection of CO 2 by the cpA neuron also activates attraction to other cues like skin odorants, visual cues, and importantly to body-warmth of 37 °C which is one of the strongest attraction cues at close-quarters 12,16,17 . The maxillary palp neuron, cpA also plays an important role in detection of human skin odorants 6 , and inhibitors such ethyl pyruvate and spermidine show reduction in attraction to skin odorants in Aedes 6,18 . Odorants like these www.nature.com/scientificreports www.nature.com/scientificreports/ could be useful in reducing host-seeking behavior and transmission of malaria, especially when used alongside others that block receptors detecting skin odorants.
Another approach to modulate behavior is using strong prolonged activators of cpA neurons that make the neuron unresponsive to CO 2 as has been shown with 2,3-butanedione and blends on An. coluzzii and Ae. aegypti 13 . However, at the higher concentration needed for this effect the unpleasant smell of this odorant and health concerns precluded integration into practical solutions. We were able to demonstrate that An. coluzzii showed an ultra-prolonged activation to (E)-2-methylbut-2-enal, which resulted in masking the detection of CO 2 significantly for several minutes after by the maxillary palp cpA neurons suggesting that this odor could disrupt detection of CO 2 and navigation toward its source as has been shown in Ae. aegypti 13 and could be utilized for potential practical applications in preventing mosquito bites and spreading of mosquito-borne diseases 19 .
Some of the CO 2 receptor neuron inhibitors have organoleptic and physicochemical properties that are conducive to development into spatial and short-range masking agents for anopheline mosquitoes. However, ultimately for a masking strategy to work, additional odorants will be needed in a blend to block other human skin receptors that are members of the Odorant receptor (Or) and Ionotropic receptor (Ir) gene families.

experimental procedures
Mosquitoes. The M form of Anopheles gambiae (Herein An. coluzzii, Ngousso strain, Cameroon) were maintained in a 12:12 (Light:Dark) photocycle at 27 °C and 70% RH. The An. sinensis mosquitoes were received from MR4 center and then colony was maintained in insectary at the same conditions as An. coluzzii. Adult females were fed on bovine blood through a heated membrane feeding system (Hemostat Laboratories, California, USA). electrophysiology. Single-sensillum recording was carried out with 4-12 days old female anopheline mosquitoes as described elsewhere 6,13,20 . All recording replicates were performed in different specimens. All odorants were obtained from Sigma at >98% purity and were diluted to (as indicated) in paraffin oil or water. A filter paper with 50 μl of the solution was inserted into a Pasteur pipette cartridge and the headspace was injected into a humidified airstream to further dilute it 3-fold as done previously 6 .
For the ultra-prolonged activators (E)-2-methylbut-2-enal, 3-methylbut-2-enal and 3-methylbutanal were dissolved at 10 −1 in paraffin oil or water, from which 50 μl of the solution was added on a filter paper inside a Pasteur pipette and the headspace was used for odor delivery as indicated in the section above. The odor delivery system was modified as shown in 6 ; solvent responses during the same recording session were subtracted. A controlled 3-s stimulus of solvent/stimulus was delivered from a Pasteur pipette into the carrier airstream. Subsequent 1-s of CO 2 (0.15%) stimuli was delivered using a MNJ-D microinjector (Tritech Research). Activity was calculated by subtracting baseline activity 1-s prior to each stimulus. Spike counting was done with Clampfit 10.3. phylogenetic analysis. The amino acid sequences of the CO 2 receptor orthologs of Drosophila melanogaster, An. coluzzii, An. sinensis, Ae. aegypti, and Cx. quinquefasciatus (signal peptide removed) were aligned with the ClustalW software, and the phylogenetic tree was constructed with the MEGA6 21 software, using the Maximum Likelihood method and LG + G matrix-based model 22 . Reliability of the branches was inferred by 1,000 bootstrap replicates 23 . sequence access numbers. The GenBank access numbers for the CO 2 receptor are as followed: An. coluzzii GR22 (XP_319142), GR23 (XP_312786), and GR24 (ABK97614); An. sinensis GR22 (KFB40998), GR23 (KFB38231), and GR24 (KFB40736); Ae. aegypti GR1 (XP_001655150), GR2 (XP_001654839), and GR3 (XP_001660602); Cx. quinquefasciatus GR1 (XP_001848097), GR2 (XP_001848828), and GR3 (XP_001848689); and Drosophila melanogaster GR21a (ABK97615) and GR63a (ABK97613).