Energy-state dependent responses of Anopheles gambiae (Diptera: Culicidae) to simulated bednet-protected hosts

In nature, Anopheles gambiae mosquitoes are found at various energy levels and such females must choose between seeking somatic energy from sugar sources and obtaining both somatic and gametic energy from blood hosts. We used a straight-tube olfactometer containing a simulated unobtainable blood host (human foot smell protected by a net) as well as a sugar source (honey odor). We assessed female probing rate and residence time at the net as a function of energy state (0, 24, 48, 72-h starved). In our trials, 0-h starved females showed low response to human odor, low probing rate, and residence time at the human odor site. By contrast, both 48 and 72-h individuals showed high response to foot odor, longer residence time, and higher probing rates. Seventy-two-h females also flew towards the honey source less often than other groups. Our findings suggest that managing sugar sources might be a viable strategy for influencing mosquito biting behavior. An. gambiae is the main vector of Plasmodium spp., the causative agent of malaria in Sub-Saharan Africa. The disease causes both mortality (∼800,000 people yearly) and morbidity in the millions, and it is a significant social-economic burden in all endemic countries (WHO 2010). To combat the spread of malaria, it is important to understand both the life-history and behavior of its vector (Chaves and Koenraadt 2010). Upon emergence, adult An. gambiae have low teneral reserves and, generally, will readily feed on sugar sources (Stone et al. 2011). After two days, females begin responding to human odor stimuli (Foster and Takken 2004) and are faced with two possible feeding choices which are often mutually exclusive due to the mosquito's size-constrained crop (Fernandes and Briegel 2004). An individual female mosquito may choose to continue feeding on readily available sugar to support somatic function, or feed from a blood host. The latter would provide energy for both somatic maintenance and reproduction but is less readily obtainable and more dangerous to acquire. Sugar availability has been linked to longer lifespan (Straif and Beier 1996, Gary and Foster 2001, Gary and Foster 2004), increased insemination rates by males (Stone et al. 2009), and is relevant to management of malaria (Gu et al. 2011). To maximize lifetime fitness, the choice between somatic and gametic energy sources (Stone et al. 2011) may be mediated by both external and internal factors (Ma and Roitberg 2008, Roitberg and Mangel 2010). External factors, such as the risk of death due to a defensive host, presence of potential predators, and a bednet may lower the mosquito's attack persistence when attempting to obtain a blood meal from a particular host (Walker and Edman 1985). Likewise, internal factors such as energy level, age, and life expectancy of the individual may affect this persistence (Roitberg et al. 1993). Both external and internal factors may be assessed by the female prior to seeking a source of energy. For years, insecticide-treated bednets (ITNs) have been the most effective tools against Anopheles bites (D'Alessandro 2001). While many studies have investigated mosquito interactions with bednets (Lines et al. 1987, Takken 2002), to our knowledge, no study has tested how energy state affects these interactions, although theoretical studies have examined this relationship (Ma and Roitberg 2008, Roitberg and Mangel 2010). Given that adults have limited ability to maintain energy at constant high levels, our interest lies in understanding the point at which a mosquito gives up on a potential blood meal and how energy state affects the probing rate and time spent with the potential host as well as the kind of interactions with an unobtainable host. Similar studies have addressed this issue (Walker and Edman 1985, Nasci 1991 with Aedes aegypti and Roitberg et al. 2010a with An. gambiae), but the mosquitoes were being actively dislocated from the host and no alternative food source was provided. In this study, we predicted that energy-deprived individuals would have increased probing rates towards an unobtainable blood host, giving up less readily than mosquitoes whose energy levels are high and thus have lower risk of starvation. In contrast, individuals with a higher energy level would have a wider range of options: they could wait for the host to become available or move on to another source of energy, seeking a more obtainable blood host or sugar source. To test these predictions, we employed a straight-tube olfactometer in which we simulated a human that is being protected by a bednet within a domicile. We tested individual female mosquitoes at various levels of starvation and gave them two potential food options: an unobtainable blood meal and a sugar source. Anopheles gambiae s.s. was reared from a colony originating from Njagi, Tanzania, in 1997. Our laboratory culture was kept at 28 ± 2° C, 83 ± 3% RH under the L:D 12:12 h photoperiod with 1 h of twilight transition (Conviron). Eggs were collected on moist filter paper from ovipositing adults that were blood-fed by one of the authors (SZ). Eggs hatched within 72 h and 1st instar larvae were placed in large white plastic containers (30 cm × 45 cm × 6 cm) with 3.5 liters of water. The larvae were fed fish flake food (Nutrafin – Hagen) ad lib until pupae were transferred in water-filled containers to Perspex cages with screening on three sides (30 cm × 30 cm × 30 cm) under the same RH and L:D conditions. Following eclosion, adults were provided with 5% sucrose solution ad lib from a medical grade cotton wick (Richmond Dental) contained in an Erlenmeyer flask. All 145 experimental females were large (wing length of 2.70–3.50 mm), 3–9 days post-eclosion, not previously blood-fed, presumed mated (mating was observed inside the cages), and without visibly swollen abdomens to increase the likelihood of female response to odor stimuli (Fernandes and Briegel 2004). We used a large, custom-made glass tube (olfactometer) (Figure 1), 145 cm in length with diameter of 17 cm which allowed unconstrained flight of the mosquito. An air pump (Petcetera AP5000 double type) pushed ambient air through an activated carbon filter at 0.025 liters/s and funnelled it to the tube at mid-height at one end of the chamber. A cotton ball containing approximately 5 g of honey (Kidd bros. unpasteurized alfalfa and clover honey), which was found to be attractive at full strength (pers. obs.), was placed on a metallic tray suspended midway into the tube. A semi-rigid, plastic, cylindrical, tan colored net (10 cm in height, 2.5 cm in diameter, 1×1 mm mesh holes), which served as a simulated bednet containing a conditioned nylon sock (Secret brand, knee high), was placed downwind from the honey, at a distance of 73 cm and was suspended from the top (Figure 1). The sock was worn by SZ for two consecutive days to condition it to be attractive to female mosquitoes (Braks and Takken 1999, Smallegange et al. 2010); on the second day, SZ performed a strenuous sport. The sock was then placed in a sealed plastic bag and incubated in the experimental chamber for one full day prior to being used in the olfactometer. The sock was substituted every four to five days and it was randomly assigned to the different starvation treatments. A vinyl pipe was placed at the upper extremity of the simulated bednet to allow the experimenter to exhale to simulate the presence of a live blood host and stimulate host-searching by the mosquitoes (Healy and Copland 1995). At the end of the tube (60 cm from the simulated bednet) we placed a plastic release chamber with a movable metallic screen. It is important to note that the experiment was arranged such that released mosquitoes encountered the simulated blood host prior to the simulated nectar source in a linear arrangement. The experimental runs were observed under dim red light (average of 1.79 μmol/(s*m2)). The olfactometer was aerated and flushed daily with plain water to minimize residual odors. Experimental olfactometer setup. A. Mosquito release chamber. B. Simulated bed net containing a conditioned sock and a vinyl tube through which the experiment exhaled air. C. Position of honey attractant. D. Carbon filter. E. Air pump. Treatments of unstarved (0-h) and 24, 48, and 72-h starvation were randomly assigned to adults aged three to nine days post-pupal-eclosion to avoid any age effect. The sugar source was then removed and substituted with deionized water (Knols et al. 1994). Female mosquitoes were randomly selected and individually placed into a release chamber. After a 3-min resting period, the wire screen of the release chamber was opened, allowing the mosquito into the olfactometer. The experimenter then began timing the start of the run and started blowing quick pulses of air every 6–8 s via the vinyl tube, which were sustained until the end of the experimental run. If the mosquito was unresponsive for 3 min the experimenter would gently tap on the walls of the release chamber until the female exited into the olfactometer at which point the experiment started. We monitored the female's probing rate by counting the number of attempted landings/probes on the sock and net (Nasci 1991) per unit time. We considered a landing to be completed once the female touched the net and did not fly more than 5 cm away from the net. We also recorded the residence time at the net with sock calculated as the time spent in physical contact with the net. The experiment was terminated if the female flew upwind past a line 5 cm in front of the cotton ball containing honey or rested on the sides of the glass tube for more than five consecutive minutes. Once a replicate was concluded, the female was removed, dried, and its wings removed and measured using an ocular micrometer and the software Analyzing Digital Images, Version 11–2008. All statistics and graphs were processed using JMP 9 and GraphPad Prism 5. We used a χ2 test to assess significance for the number of An. gambiae responding to the human odor stimulus as well as the number of mosquitoes that actively flew to the honey source or rested. Kruskall-Wallis test was used for analyzing significant differences in the probing rate, residence time at sock and net, as well as the average time per probing event between treatments, due to the zero inflated distribution of the values. Lastly, Log-Rank (Mantel-Cox) (i.e., survival analysis) test was used to analyze the differences in sustained activity towards obtaining a blood meal as a function of energy state. Age stratification (number of days after emergence) was included in each analysis. A total replicate size of 145 females met our criteria of evaluation. Single experimental runs lasted from a few seconds to a maximum of 27 min (average ∼3 min). Non-starved mosquitoes (0-h) responded to the human odor by probing the simulated bednet significantly fewer times than all the other treatments, as only 43% of the individuals were recorded to have probed at least once during the experiment. The other groups had a higher proportion of individuals responding to the human odor at 71%, 74%, and 67% for 24, 48, and 72 h, respectively (χ2 df.= 3, F= 25.91, p<0.001) (Figure 2). Proportion of An. gambiae mosquitoes responding to human odor by attempting to probe the simulated bednet as a function of energy state. Individuals that were not starved (0 h) showed a significantly lower propensity to respond to human foot odor (χ2, p<0.001). Different letters indicate significant differences. The relative residence time at the sock and net (i.e., the proportion of experimental time spent probing) of non-starved females was 22±0.1% and was significantly different from the 48-h and 72-h groups (Kruskall-Wallis F= 15.45, p= 0.002). These treatments yielded to 50±0.1% and 47±0.1% relative residence time, respectively, while the 24-h starvation treatment averaged at 37±0.1% (Figure 3). Mean residence time at the sock and net of female An. gambiae as a function of energy state. Mosquitoes that were more energy depleted (48, 72 h) spent a significantly longer portion of the experimental time probing at the bednet (Kruskall-Wallis, p=0.002). Error bars show bootstrapped 95% CI. Different letters indicate significant differences between bars. When considering the proportion of individuals that actively left the “hut” by flying to the honey source at one end of the olfactometer vs the proportion that rested in the immediate surroundings of the human odor, the only significantly different group was the 72-h treatment in which 24% of the individuals moved to the honey (χ2, df.= 3, F= 35.02, p<0.001). The other groups scored 48%, 51%, and 65% for 0, 24, and 48 h, respectively (Figure 4). Proportion of An. gambiae that actively flew towards a honey source placed at one end of a single arm olfactometer (Figure 1) or rested for at least 5 consecutive min in proximity of the human foot odor source. Individuals in the 72 h food deprivation treatment flew towards the honey significantly less (χ2, p=0.008). Different letters indicate significant differences. The proportion of females still active in host-seeking after a given time, measured as total experimental time, yielded significant differences between the two extreme treatments: 0 and 72 h. The latter showed a much longer overall active host-seeking time length (Log-Rank Mantell-Cox F= 7.985, p=0.046) (Figure 5). Kaplan-Meier survival plot of An. gambiae activity time (i.e., proportion of active individuals not flying towards honey scent and not resting for 5 consecutive min) as a function of energy state. Very starved individuals (72 h) stayed significantly more active over time around the bednet than non-starved ones (0 h) (Log-Rank Mantell-Cox, p=0.046). Probing rate (total probes/total experimental time) was significantly different between 0-h (0.239±0.001 bites/min) and 48-h (0.481±0.001 bites/min) treatments (Kruskal-Wallis, F= 9.563, p=0.023). 24 h and 72 h were recorded at 0.407±0.001 and 0.430±0.001 bites/min, respectively. No significant results were obtained when comparing the average length of time per landing (∼67 s across all levels) on the simulated bednet between energy levels (Kruskall-Wallis F= 4.232, p=0.238). In this investigation, mosquitoes with different energy levels were offered a choice between an unobtainable (simulated) host under a bednet and an alternative source of energy, honey, a proxy for a nectar source outside of a domicile. The experiment was designed spatially such that the mosquito would encounter the blood host before the honey (Figure 1). The goal was to see if females would seek the unobtainable blood meal indefinitely or give up this potential source of energy to actively fly towards the alternative sugar source. We found that An. gambiae female mosquitoes modulate their behavior toward trying to bite an unobtainable host based on their energy state. In nature, females are found with a broad range of energy states, and our research helps elucidate how they might react to the presence of two potential energy sources with very different fitness advantages and costs for the individual female. Mosquitoes that were granted ad libitum access to sugar water until the start of the experiment (0 h starved) were overall less inclined to respond to any of the two odor stimuli (Figure 2). This absence of response suggests that the high energy level, perhaps coupled with a perceived full crop, reduces meal-seeking activity (Jones and Madhukar 1976). These findings are in line with Roitberg et al. (2010a) where it was shown that well-fed females had lower attack rates and were predicted by Ma and Roitberg (2008). In the few instances where these females probed for blood, their residence time at the sock and net was short and they tended to have lower probing rates and to terminate blood-host-seeking significantly earlier than other groups (Figures 3, 4, and 5). Here, we suggest that recently fed females may postpone the risk of attacking a blood host or venturing away from it until crop volume and mass drops, perhaps maximizing the size of the future engorged meal and survivorship (Roitberg and Gordon 2005, Ma and Roitberg 2008). This can be best understood using a marginal analysis wherein the marginal returns from both space-constrained blood meals and sugar meals are small for recently fed individuals, whereas the marginal costs via mortality risk are high (see Roitberg et al. 2010b for an example from parasitic wasps). In contrast, the marginal returns for starved individuals are high in either case, but one of the two meal options might be more accessible. This statement is supported by our data. Very starved females (48 and 72 h) had longer residence times at the sock and net than other groups and 48 h females had a significantly higher probing rate. However, 72 h starved females showed no significant difference in probing rate. Perhaps females in this category are on the verge of starvation and are energetically constrained to overall low activity levels. This hypothesis is supported by the extremely high mortality in the 72 h treatment cages and the fact that this treatment yielded females that rested alongside the simulated blood-host more often than other groups; we hypothesize that the latter had sufficient energy to fly towards the sugar source in search of alternative food. However, we cannot exclude the possibility that mosquitoes that flew towards the honey source were simply flying upwind. Although some females were observed to readily feed on the cotton bearing honey, no data were collected regarding honey avidity. Finally, the low probing rate and residence time at the sock and net by well-fed individuals suggests that the female's energy level might allow for a flight to another perhaps more readily accessible blood host (Lines et al. 1987). Very starved individuals (48 h, 72 h treatments) not only responded more often to odor stimuli but also showed a higher persistence rate and residence time at the sock and net. The propensity to stay in contact for a longer duration of time with the bednet could be advantageous in the field use of insecticide treated bednets (ITNs), as increased contact time with the chemicals proportionally increases the likelihood of death by poisoning (Roitberg and Mangel 2010). The lack of sugar sources in the immediate vicinity of human habitations may contribute to the overall depletion of energy in females, increasing the exposure to the chemicals on ITNs, and this fact should be considered by malaria vector managers. Similar considerations are relevant for females that have undergone changes due to Plasmodium infection (shortened overall life expectancy and increased risky behavior) as they may be linked to depletion of energy in the mosquito host (Ferguson and Reid 2002, Rivero and Ferguson 2003). We were expecting that starved individuals would spend on average a longer time per probe but that did not occur. There are at least two explanations for this finding. First, An. gambiae could have evolved an maximum length of time allowed per probe (see giving up times in and which could have evolved in response to the risk of by the host and potential death of the mosquito. by an feeding the energy state at the unobtainable host was starved and well-fed individuals gave up at the same time but for different with starved individuals immediate to energy state and well-fed individuals risky for future (see Roitberg and Mangel 2010). Our findings suggest that the management of sugar sources within a should be a In as by et al. the then management of within to in a of We the management to within in management and the of sugar which have shown to be an strategy in and 2008). research is in this field and we are An. gambiae in the same in conditions. We and for their in this experiment. to and for the support and the Roitberg for and for of this

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.