This study examined Culex pipiens pallens responses to different combinations of colors and chemicals employed via a mosquito trap under semifield conditions. Our results indicated that Cx. p. pallens has color and chemical concentration preferences. Culex p. pallens had a 38.0% greater response to white than black color treated traps. Further, Cx. p. pallens showed differences in olfactory attraction depending on the chemical and concentration. Culex p. pallens was 107.6% more attracted to traps employing 500 ppm ammonia than control (i.e., unscented). Similarly, Cx. p. pallens was 117.5%, 128.8%, and 140.3% more attracted to traps employing, respectively, 1,000, 10,000, and 20,000 ppm of ammonia hydrogen carbonate compared to controls. And the response to lactic acid showed that Cx. p. pallens was most attracted to concentrations of 100 and 500 ppm (135.7% and 142.9%, respectively) compared to controls.
Pathogen transmissions from hematophagous arthropods account for ∼17% of all infectious diseases, resulting in more than 700,000 deaths annually (WHO 2017) and enormous economic losses through costs of vaccinations, vector controls, and trade embargoes; e.g., the Zika virus outbreak in Latin America cost ∼$18 billion from 2015 to 2017 (Osburn 2008). Mosquitoes are known for vectors transmitting multiple pathogens causing infectious diseases that adversely affect both humans and animals (WHO 2017). Humans have inadvertently helped mosquitoes spread such pathogens. For instance, rapid advances in global transportation have benefited humanity. However, these advances also facilitate the spread of mosquito populations globally, which furthers the spread of arboviruses such as chikungunya, yellow fever, and Zika (Tatem et al. 2006). Rising global temperatures are helping expand mosquito distributions, and warmer temperatures are reducing their extrinsic incubation periods of pathogens (Boukraa 2016, Samuel et al. 2016). These synergistic effects make mosquito distribution and pathogen transmission difficult to predict. Therefore, the development of novel vector-control and surveillance strategies, such as effective mosquito traps, has been identified as a top priority by the World Health Organization (WHO 2017).
Female Culex p. pallens (Coquillett) feeds primarily on protein-rich (69–97%) avian blood (Farajollahi et al. 2011) but will also feed on mammalian blood to obtain the protein levels required for egg development (Gómez-Díaz and Figuerola 2010). Although Cx. p. pallens is an opportunistic mammalian blood feeder, this species does play a principal role in the transmission of a wide range of zoonotic pathogens affecting humans, livestock, pets, and wildlife including but not limited to West Nile virus and St. Louis encephalitis virus, Rift valley fever virus, and filarial worms (Barnard and Xue 2004, Kim and Tsuda 2010). Furthermore, Cx. p. pallens is a highly adaptive species that thrives on anthropogenic disturbances in urban environments. Culex p. pallens utilizes a wide range of breeding sites, such as man-made ponds, ditches, sewage ways, septic tanks, and artificial containers (e.g., livestock drinking tanks, unattended birdbaths, or plastic pet pools) around farms and households. Consequently, these urban environments allow for additional and continuous (e.g., winter) breeding habitats and proliferate food resources (e.g., algae and other micro-organisms) for mosquito larvae without the need for annual diapause (Savage and Miller 1995). Ultimately, with the ability to feed on multiple hosts, a high vectorial capacity, and the affinity for environmental plasticity, Cx. p. pallens is a threat to the health of many organisms, including humans (Vinogradova 2000). However, little is known about the determinants of host choice in Cx. p. pallens.
Mosquitoes, including Cx. p. pallens, evolved successful strategies for locating appropriate hosts over short distances through both optical and olfactory sensory pathways (Michalet et al. 2019). Mosquitoes have dichromatic vision with specific wavelength discrimination abilities (Muir et al. 1992, Peach et al. 2019), allowing them to respond to color, contrast, and light intensisty when seeking a host (Allan et al. 1987). Similarly, specific odorant receptors exist on the mosquito antennae (Carey et al. 2010), maxillary palps (Athrey et al. 2017), and tarsi (Bentley and Day 1989), which discern odors, typically in the form of volatile organic compounds, associated with an appropriate host (Lemasson et al. 1997). However, host odor concentrations must be within a specific range to be both detected and attractive, and certain conentrations can even act as a deterrent (Wheatley 2002).
While much progress on mosquito host attractiveness has been made, much of the work has focused on diurnal and anthropophilic mosquito species (e.g., Anopheles gambiae Giles and Aedes aegypti [L.]). Therefore, we investigated both color and chemical attraction in Cx. p. pallens and asked 2 questions: 1) How do the surface colors and patterns of a mosquito trap affect attraction? and 2) How do different chemicals and their associated concentrations affect mosquito attraction to a trap? We investigated these questions under semifield conditions by exposing female Cx. p. pallens to either 3 different colored surfaces (white, black, or a checkered pattern) or different concentrations of 3 different human-oriented chemicals (ammonia, ammonium hydrogen carbonate, and lactic acid).
MATERIALS AND METHODS
A Culex p. pallens colony was maintained in the medical entomology laboratory (27.0 ± 1.0°C, 60.0 ± 10.0% RH, and 14:10 light and dark photoperiod) at Kosin University, Busan, Republic of Korea. The larvae (∼1,000) were reared in an enamel rectangular pan (50 cm × 40 cm × 10 cm) containing tap water. Larvae were fed a 2:1:0.5 mixture of beef chow, tetramine fish food, and brewer's yeast daily (Gerberg et al. 1994). Each day, pupae were collected and placed in a 500 ml glass beaker of tap water at a density of 100 pupae/beaker and placed into 45 cm × 45 cm × 35 cm aluminum screened wire mesh cages (BioQuip Products Inc., Rancho Domínguez, California) for adult eclosion. Emergent adults were provided ad libitum with a 10% sucrose solution placed on absorbent cotton rolled in cotton-muslin gauze cloth and inserted in a 50 ml glass bottle placed inside each adult cage. A week after emergence, female mosquitoes were offered live rats (Rattus norvegicus Berkenhout) to feed on. Forty-eight hours after blood-feeding, a 500 ml glass beaker containing tap water was provided as an oviposition site in each cage. Females were allowed to deposit eggs in the beaker for 3 days, after which the eggs were transferred to an enamel rectangular pan to hatch.
Light-emitting diode mosquito traps (L-trap; Eregreen Co., Gwangju, Korea) were used in this study (Fig. 1). Each trap consisted of 12 small light-emitting diode lamps on a plastic trap body 20 cm × 20 cm × 0.7 cm (Table 1). Each trap had 7 fans powered by an electric 8-V AC motor that drew air through an opening in the front of the trap and exhausted it through the back of the trap. This wind current sucked mosquitoes inside and stored them in a collection chamber.
For each trial, 20 or 50 laboratory-raised and mated female mosquitoes (5–8 day-old) that had never been offered a blood meal were collected into a sterile cylindrical tube (diam 8 cm, length 10 cm), using a battery-powered aspirator (Hausherrs Machine Works Co., Toms River, New Jersey). Mosquitoes were released into a 1.8 m × 3.7 m × 1.8 m flight-screened enclosure in dark conditions (approximately 50 lux) 10 h prior to each trial to become familiar with environmental conditions (27.0 ± 1.0°C and 60.0 ± 10.0% RH). Experiments were initiated 30 min before the end of the light cycle (12:12 L:D), corresponding to the nocturnal feeding activity of Cx. p. pallens (Veronesi et al. 2012).
During experiments, we remained outside the room to avoid contaminating the study with our own bodily volatile emissions. Traps were assigned by a random number generator and rotated clockwise between trials to prevent positional bias. Additionally, traps were cleaned with 70% ethanol between trials to further prevent any chemical bias.
Visual response assay:
Visual response tests were designed to evaluate the collection efficacy of an L-trap with 1 of 3 different colored attractants: solid black, solid white, or a black-and-white checkered pattern. The front of the trap was covered with polyethylene sheets of white, black, or a checkered pattern (Fig. 1) that were attached with a scentless paste. Fifty females Cx. p. pallens were released into the flight-screened enclosure and exposed to a single colored trap for 14 h following their 10 h exploration period (described above). Each color trial was replicated 3 times resulting in a total of 450 mosquitoes (n = 150 per color treatment). Attractiveness was defined as the total number of mosquitoes captured in a collecting chamber after each trial. Collections during trials took place at 1, 2, 4, and 14 h to further assess if attraction changed over time. To minimize disturbance of behavior, the collection of mosquitoes during experiments lasted less than 2 min. Following each trial, the testing enclosure was cleared of any remaining mosquitoes and wiped down to remove chemical buildup from mosquitoes during experiments. A fresh group of mosquitoes was used in all tests.
Olfactory response assay:
The olfactory attractant tests were designed to evaluate the collection efficacy of an L-trap baited with 1 of 3 different chemical attractants, each at 5 different concentrations; ammonia solution, ammonium hydrogen carbonate, and lactic acid each at 100, 500, 1,000, 10,000, and 20,000 ppm. The LED-light source on the L-trap was removed for olfactory response assays to focus solely on chemical attraction. Ammonia (28–30%; Junsei Chemical, Tokyo, Japan) was dissolved in 85.6 ml of distilled water at 50°C. Then 10.1 g of gelatin powder (Kanto Chemical Co., Tokyo, Japan) was added to the mixture. When the solution solidified, 4 ml of ethyl alcohol (94%; SK-Chemicals, Ulsan, Korea) and 100 μl of dipropylene glycol (100%; What Soap Co., Cheong Ju, Korea) were added as volatile enhancers. To prevent the proliferation of mold, 100 μl of grape seed extract (70%; What Soap Co, Cheong Ju, Korea) was added to the solutions and vortexed to homogenize. After homogenization, the gelatin compound was transferred to a plastic storage container (5.3 cm × 2.7 cm × 9.8 cm) with 4 holes (diam 4 cm) at the bottom to facilitate volatile emissions. The same method was used for ammonium hydrogen carbonate (95%; Junsei Chemical, Tokyo, Japan) and lactic acid (30%; Sigma, St. Louis, Missouri) attractants.
For the olfactory response assays, 20 females were released into the flight-screened enclosure set to the same environmental conditions (27.0 ± 1.0°C and 60.0 ± 10.0% RH) used for the visual response assays. During experiments, an attractant at a given concentration was placed in a plastic container inside the compartment located at the top of the trap. Two traps were placed parallel to one another on opposite ends of the cage, 1 with a chemical attractant (i.e., treatment) and 1 with water (i.e., control). Each chemical attractant and concentration were replicated 3 times (n = 60 per chemical treatment). For each replicate, 20 females were released, and the total numbers of mosquitoes found in the trap collection chamber of both the treatment and control traps were quantified following a 14 h assay period.
For the visual response assays, the effect of trap color (white, black, or checkered pattern) across the exposure period was determined by a 1-way ANOVA. For the olfactory response assays, a paired t-test was used to compare mosquito responses between treatment and control across chemical concentrations using SPSS Statistics for Windows, Version 13.0 (IBM Corp., New York, NY). For post hoc comparisons of both the visual and olfactory response assays, an independent Duncan's multiple range test was used for comparisons of means but detransformed for the figures in the result section. Alpha for all comparisons was set at 0.05.
Overall, solid white and checked traps had a greater trapping efficiency (i.e., percent mosquitoes captured) compared to black traps (98.0% and 96.0% vs. 66.7%, respectively; Fig. 2). When considering the main effects of time (P < 0.0001), trapping efficiency increased with time, regardless of color (Fig. 2). However, at certain time intervals, trapping efficiency differed by trap color. At 1 h of exposure, more mosquitoes were trapped by white traps than either black or checkered patterned traps (P < 0.0129 or P < 0.0067). And at both 4 and 14 h, both white and checkered traps had significantly greater trapping efficiencies (P < 0.0055 and P < 0.035) than black traps. While the main effects of trap color (white, black, or checkered pattern) and exposure time (1, 2, 4, or 14 h) (P < 0.0001) significantly impacted mosquito response, there was no significant treatment by time interaction (P < 0.1072).
When mosquitoes were exposed to traps emitting ammonia, trap efficiency was significantly greater than controls only at a concentration of 500 ppm (33.3% vs. 10.0%; P < 0.001; Fig. 3). Ammonia with greater than 1,000 ppm concentration elicited a lower number of responses by mosquitoes compared to traps with 100 and 500 ppm concentrations and controls (Fig. 3). Traps emitting ammonia hydrogen carbonate had significantly higher trapping efficiencies (P < 0.001) at concentrations of 1,000 ppm (45.0% vs. 11.7%), 10,000 ppm (38.3% vs. 8.3%), and 20,000 ppm (66.7% vs. 11.7%) compared to controls (Fig. 4). An L-trap with an ammonia hydrogen carbonate concentration of 20,000 ppm collected 56.7%, 43.4%, 21.7%, and 28.4% more Cx. p. pallens than traps emitting 100, 500, 1,000, and 10,000 ppm concentrations of ammonia hydrogen carbonate for a 14-h trial, respectively (Fig. 4). Traps emitting lactic acid had a greater trapping efficiency compared to controls at all concentrations, though statistical significance (P < 0.001) was observed only at 100 ppm (35.0% vs. 6.7%) and 500 ppm (30.0% vs. 5.0%) concentrations compared to controls (Fig. 5). Although lactic acid at greater than 1,000 ppm concentrations showed no significant difference in the total number of mosquitoes attracted in relation to control, the collection number of the 10,000 ppm group was 30.0% greater than other concentrations.
Both visual and olfactory stimuli function as a cue and a possible signal for mosquitoes with regard to locating hosts. We found Cx. p. pallens to be more visually attracted to a white mosquito trap (or black-and-white checkered pattern) than to a solid black trap (Fig. 2). Similarly, we found Cx. p. pallens were more attracted to traps emitting human-oriented volatiles compared to controls (Figs. 3–5). However, the level of attraction was dependent on the concentration of the volatile emission. These results provide an evidence of the behavioral shifts in the Cx. p. pallens host-seeking behavior to both visual and olfactory stimuli, which ought to be considered when developing mosquito traps.
Although mosquito responses to visual stimuli vary across species, most diurnal mosquitoes prefer dark (e.g., black, checkered) over lighter (e.g., white) colors (Hecht and Hernandez-Corzo 1963). Consequently, most of the mosquito traps used for population sampling are dark cylinder suction traps. However, we found the nocturnal Cx. p. pallens selected white traps more than black traps in a simulated nocturnal setting. Thus, the effectiveness of trap color may differ by activity time of mosquitoes. For instance, Muir et al. (1992) demonstrated that mosquitoes have the physiological ability to sense both light intensity and spectrum. Furthermore, Silva et al. (2005) found that mosquito response varied with reflected and transmitted color. Thus, 1 possible explanation as for why Cx. p. pallens was more attracted to white traps than black traps could be related to light intensity and reflectiveness. Because white is more reflective than black, the white traps may have been more visible to Cx. p. pallens during experients, thus revealing that white traps may be more effective for capturing nocturnal mosquitoes. However, more studies with additional nocturnal species are needed to confirm this hypothesis.
In addition to visual stimuli, mosquitoes are attracted and repelled by specific chemical compounds (Takken 1991, Logan 2008). Skin microbes have co-evolved with hosts (e.g., mammals) through strong ecological interaction (e.g., metabolite conversion, signaling, chemotaxis) and convert odorless host skin residues (i.e., sweat) to aliphatic and aromatic carboxylic acids (Bosch et al. 2000), which volatilize and serve as chemical signals to various organisms, including mosquitoes (Wooding et al. 2020). When exposed to different concentrations of ammonia, a concentration of 500 ppm resulted in the greatest attractiveness for Cx. p. pallens (Fig. 3). As described by Healy et al. (2002), a small amount of volatile host-emitted compounds, such as carboxylic acids, are critical to the mosquito's ability to discriminate against a nonpreferred host at close range. In our study, all levels of emitted ammonia (100, 500, 1,000, 10,000 or 20,000 ppm) resulted in greater trap efficiency compared to controls (18.3%, 20.3%, 13.4%, 3.3%, and 13.3%). A study by Delventhal et al. (2017) showed there is an olfactory trend for chemicals that induce opposing responses through the same sensory system when presented in higher concentrations. For example, in alignment with Smallegange et al. (2005), ammonia concentrations exceeding 500 ppm were not as attractive to mosquitoes and consequently reduced the trap efficiency to be no better than controls. Moum et al. (1969) found humans emited ammonia at ∼25 to 100 ppm for several hours, which is within the range we found for high levels of mosquito attraction, though we did not test levels below 100 ppm. Mosquitoes may recognize low concentrations of ammonia as a cue for the availability of a food source at a distance, whereas high concentrations sensed through gustation warn of potential toxicity (Delventhal et al. 2017). Thus, ammonia might elicit an aversive olfactory response and indicate poor host quality (or toxic) at higher concentrations. Although mosquitoes were still generally more attracted to the higher levels of ammonia compared to controls in our study (Fig. 3), they likely use a combination of olfactory and visual cues when deciding on host quality. Thus, any level of ammonia may initially attract a mosquito to a host, but the mosquito may not feed.
Ammonia hydrogen carbonate, which is a source of ammonia emanation, has been used as an attractant for a variety of insects, including cherry fruit flies (Diptera: Tephritidae) (Kendra et al. 2009), blowflies (Diptera: Calliphoridae) (Cragg and Thurston 1950), and mosquitoes (Anderson et al. 2012). In our study, the traps emitting 1,000 ppm, 10,000 ppm, or 20,000 ppm ammonium hydrogen carbonate significantly increased the host-seeking behavior of Cx. p. pallens compared to control traps (Fig. 4). These results suggest that the attractiveness of ammonium hydrogen carbonate is not effective unless at relatively high concentrations (1,000 ppm or greater). However, ammonium hydrogen carbonate has the unique chemical property of generating gaseous ammonia and carbon dioxide upon heating, which are additional attractants for mosquitoes (Zermoglio et al. 2017). Therefore, ammonium hydrogen should not be discounted as an attractant for Cx. p. pallens, but additional testing incorporating different temperature ranges to validate the true effectiveness of ammonium hydrogen volatility resulting in mosquito attractiveness is needed. This would offer the possibility of the use of a mosquito trap with a simple heated device to both increase the volatility of ammonium hydrogen and provide further thermal cues to mosquitoes.
Lactic acid is 10–100 times more abundant on human skin than on other animals (Geier et al. 1996, Dekker et al. 2002) and considered to be an effective chemical compound attracting anthropophilic mosquito species (Dekker et al. 2002). However, there are very few studies that investigated lactic acid preference with Cx. p. pallens. Our results revealed that lower concentrations (100 and 500 ppm) of lactic acid induced Cx. p. pallens host-seeking behavior and consequently led to greater trapping efficiency (Fig. 5). It is not surprising that Cx. p. pallens was attracted to lower concentrations of lactic acid, as lactic acid is present in low quantities on avian skin (Dekker et al. 2002), a primary host of Cx. p. pallens. Thus, higher concentration may not be identified as a strong attractant to aviphilic Cx. p. pallens. Additionally, it is important to note that birds do not have sweat glands (Rath et al. 2015). As described above (Bosch et al. 2000, Braga et al. 2016), human host skin residues (i.e., sweat) play an important role in mosquito perceiving and responding to external information as a cue from bacteria metabolic products.
Sweat is not only directly interacting with skin bacteria but also facilitates evaporation at ambient (or above ambient) temperatures. Furthermore, because mosquitoes are not stationary, their responses to volatile chemical compounds vary depending on the direction and concentration of volatile molecules in the air with respect to the turbulent flow (Geier et al. 1999), as well as the mosquitoes' location. Additional variables will be needed in order to determine whether host body size and physical status such as humidity (i.e., sweat) changing the plume structure influence mosquitoes' host-seeking behavior.
Culex p. pallens are well adapted to urban ecosystems, utilizing artificial bodies of water (e.g., septic tanks) for both breeding and overwintering (Nelms et al. 2013). Consequently, mosquito hosts, including humans, are at higher risk of exposure to mosquitoes and ultimately to mosquito-borne diseases. Various mosquito traps have been developed and used to kill and monitor mosquito populations, as well as to track infection rates of human and animal mosquito-borne pathogens (Curtis 1996). Our study determined ecologically relevant cues of both visual and olfactory sensory modes, as related to mosquito host-seeking behavior. These results demonstrated a white trap surface increased attraction. As a sequential experiment equipped with the white trap, specific ranges of chemical concentrations additively (a possible synergistic effect) increased mosquito attractions and enhanced trap efficiency. Continued improvement in the efficiency of mosquito traps based on both visual and chemical attraction is needed. Future work could investigate the interactive (or synergistic) effects of chemical compounds and trap color, as well as the interactions between different chemical compounds and chemical concentrations to reveal any potential synergism in Cx. p. pallens host-seeking behavior.
The authors thank the personnel of the Medical Entomology Laboratory at Kosin University for providing support for this study. The authors also thank the anonymous reviewers' feedback on the earlier versions of this manuscript.
Florida Medical Entomology Laboratory, University of Florida, Vero Beach, FL 32962.
Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82072.
Department of Health and Environment, Kosin University, Busan, Republic of Korea.