Development and field evaluation of a synthetic mosquito lure that is more attractive than humans

Abstract

Background: Disease transmitting mosquitoes locate humans and other blood hosts by identifying their characteristic odor profiles. Using their olfactory organs, the mosquitoes detect compounds present in human breath, sweat and skins, and use these as cues to locate and obtain blood from the humans. These odor compounds can be synthesized in vitro, then formulated to mimic humans. While some synthetic mosquito lures already exist, evidence supporting their utility is limited to laboratory settings, where long-range stimuli cannot be investigated.

Methodology and principal findings: Here we report the development and field evaluation of an odor blend consisting of known mosquito attractants namely carbon dioxide, ammonia and carboxylic acids, which was optimized at distances comparable with attractive ranges of humans to mosquitoes. Binary choice assays were conducted inside a large-cage semi-field enclosure using attractant-baited traps placed 20 m apart. This enabled high-throughput optimization of concentrations at which the individual candidate attractants needed to be added so as to obtain a blend maximally attractive to laboratory-reared An. gambiae. To determine whether wild mosquitoes would also be attracted to this synthetic odor blend and to compare it with whole humans under epidemiologically relevant conditions, field experiments were conducted inside experimental huts, where the blend was compared with 10 different adult male volunteers (20-34 years old). The blend attracted 3 to 5 times more mosquitoes than humans when the two baits were in different experimental huts (10-100 metres apart), but was equally or less attractive than humans when compared side by side within same huts.

Conclusion and significance: This highly attractive substitute for human baits might enable development of technologies for trapping mosquitoes in numbers sufficient to prevent rather than merely monitor transmission of mosquito-borne diseases.

Figure 1. A conceptual model summarizing the development of the odor blend.

The process began with a weakly attractive mixture containing 2.5% aqueous ammonia and CO₂ gas flowing at 500 ml/min, which was enhanced by adding 85% L-lactic acid (LA). Onto the resulting mixture, each of the other aliphatic carboxylic acids was added separately, each of them at their optimally attractive concentrations. The blend therefore consisted of the CO₂ gas plus hydrous solutions of ammonia (2.5%) and L-lactic acid (85%), and the other aliphatic carboxylic acids at their respective optimum concentrations as follows: propionic acid (C3) at 0.1%, butanoic acid (C4) at 1%, pentanoic acid (C5) at 0.01%, 3-methylbutanoic acid (3mC4) at 0.001%, heptanoic acid (C7) at 0.01%, octanoic acid (C8) at 0.01% and tetradecanoic acid (C14) at 0.01%. Finally, a variant of the blend was formulated by removing 3mC4.

Figure 2. Long range performance of the synthetic blend and its variant.

Average number of mosquitoes caught per night, inside experimental huts whenever the blend, its variant (without 3-methyl butanoic acid) or a human volunteer was inside the hut. There were significantly more mosquitoes caught in huts baited with the blend or its variant than in huts baited with humans (P<0.001) as determined by General Linear Model using SPSS version 15 (SPSS Inc.). However, no difference was observed between the blend and its variant (P>0.05). The error bars represent 95% confidence intervals.

Figure 3. Short range performance of the synthetic blend.

Addition of the synthetic blend into huts occupied by human volunteers significantly increased the number of mosquitoes caught in the huts compared to huts with only a human volunteer (A). However, once the mosquitoes were inside the huts, preferences for either bait type were similar for An. funestus (P = 0.179) and the other unidentified Anopheles mosquitoes (P = 0.82) but humans remained significantly more attractive to An. gambiae s.l. (P<0.001), Culex (P = 0.045) and Mansonia mosquitoes (P<0.001) (B), by General Linear Model using SPSS version 15 (SPSS Inc.). The error bars represent 95% confidence intervals.