Anopheles Gambiae s.s Attraction to Floral and Human-Skin-Based Odors and Their Combination is Modulated by Previous Bloodmeal Experience

Background: Mosquitoes use odors in nding energy resources, blood hosts and oviposition sites. While these odor sources are normally spatio-temporally segregated in a mosquito’s life history, here we explored to what extent a combination of ower- and human-mimicking synthetic volatiles would attract the malaria vector Anopheles gambiae s.s. Methodology: In the laboratory and in large (80 m 2 ) outdoor cages in Tanzania, nulliparous and parous Anopheles gambiae s.s. were offered choices between a blend of human skin volatiles (Skin Lure), a blend of oral volatiles (Vectrax), or a combination thereof. The blends consisted of odors that induce distinct, non-overlapping activation patterns in the olfactory circuitry, in sensory neurons expressing olfactory receptors (ORs) and ionotropic receptors (IRs), respectively. Catches were compared between treatments. Results: In the laboratory nulliparous and parous mosquitoes preferred skin odors and combinations thereof over oral odors. However, in semi-eld settings nulliparous were signicantly more caught with oral odors, whereas no differences were observed for parous females. Combining oral and human volatiles did not augment attractiveness. Conclusions: Nulliparous and activate meal


Background
There has been a decline of malaria incidence and prevalence globally over the past decade [1][2][3][4][5]. Several interventions have been reported to play substantial roles in this decline, including improved case management and malaria diagnostic methods, as well as the deployment of long-lasting insecticidetreated nets (LLIN) and indoor residual spraying (IRS) to reduce interactions between mosquito vectors and people [2,[6][7][8]. However, the positive results achieved through the universal coverage of LLINs and IRS have caused the community, policy makers, and other malaria control stakeholders to redirect resources toward these interventions and away from other vector control techniques [9]. Vector control campaigns, particularly those seeking to reduce transmission of malaria, have shown that over-reliance on a single approach or a certain group of insecticides leave a campaign vulnerable due to e.g. the development of resistance [10].
Repeated application of insecticides with a similar mode of action can lead to resistance development in mosquitoes [10]. In the rst global malaria eradication campaign between 1955 and 1969 intensive indoor residual spraying of dichlorodiphenyltrichloroethane (DDT), although initially very successful in pushing back malaria, eventually led to DDT resistance in mosquitoes and failure of the campaign in different areas [9,[11][12][13]. More recently, mosquito resistance to pyrethroid-based insecticides used for malaria vector control has been reported in several countries [9,[14][15][16]. In the face of increasing mosquito resistance and a threat of malaria resurgence, the World Health Organization (WHO) recommends the use of integrated vector management (IVM) [17], which employs several scienti cally proven, complementary methods of intervention to control all vectors [17]. Hence to complement current interventions and to sustain the gains of universal coverage by LLINs and IRS, further exploration of novel and innovative strategies is of paramount importance.
LLIN and IRS capitalize on mosquitoes being attracted to humans, but intercepts and kills mosquitoes before they reach the host. A complementary approach would be the use of odour sources other than humans to divert mosquitoes away from biting humans. These odors can be derived from owers and extra oral nectaries [18,19], from oviposition sites [20], or from blood hosts such as humans [21,22], and can be used in for instance interfering with mating or oviposition, or in eliminating vectors.
The latter, called attract-and-kill, involves attracting mosquitoes to odor baits laced with a toxic agent that kills the vector upon contact. By ne-tuning blends, baits can be developed that selectively target mosquitoes and have minimal impact on the environment. These have the potential to make a signi cant contribution to mosquito population management and the suppression of mosquito-borne disease [21,[23][24][25][26][27]. Such methods are increasingly employed against pest insects in agriculture and in vector control.
The attraction of mosquitoes to single compounds emanating from humans has been demonstrated several times [28][29][30], and a single plant-based compound also has been reported to be effective [31]. In addition, studies have explored possible additive or synergistic effects of blends of semiochemicals to increase mosquito attraction, either through mimicking human volatiles [32][33][34][35] or putative host plants [36]. However, initial tests of combined human and plant volatiles have been investigated only quite recently [37][38][39][40], with mixed or inconclusive results. The objective of the present study was to evaluate how combining odor blends would reduce or increase attractiveness to the major African malaria vector, Anopheles gambiae s.s. The rationale behind the choice of synthetic blends used was that these are characteristic for spatiotemporally distinct sources (vertebrate and oral). In addition, the blends induce sensory responses in distinct subsets of sensory neurons that either express ionotropic receptors (IRs, in grooved peg neurons) or neurons that express olfactory receptor, and could thus behaviorally complement (addition or synergy) or antagonize (reduced attraction) each other.

Study area
This study was carried out in Muheza District located in the northeast of Tanzania (5°13′S, 38°39′E; altitude 193 m). The district is characterized by a humid and warm climate almost throughout the year. The average annual rainfall in Muheza is 1,000 mm with two seasonal peaks i.e. a main peak between March and May, and a less pronounced one between November and December. The mean temperature in the area is 26 o C, with below-average temperatures between June and September and above-average between October and May. The experiments were carried out in an insectary and in mosquito spheres [41] at the Amani Research Centre of the National Institute for Medical Research.  [42]. European Community guidelines and standards were followed in rabbit maintenance [42]. Only female mosquitoes were used for both laboratory and semi-eld trials.
Chemical/ Odour blends used 1. Vectrax (ISCA Technologies, Riverside) is sprayable liquid formulation comprised of a synthetic mix of typical oral volatiles that mimic sugar-rich owers and extra oral nectaries from which mosquitoes of all species and both sexes seek sustenance throughout their lives [23]. Mosquitoes detect these oral attractants, released over time from the Vectrax formulation, and respond by orienting their ight toward the point source. Vectrax also contains several sugars-and protein-based feeding stimulants, which encourage mosquitoes to feed upon the formulation to full engorgement [23].
2. Skin Lure is a matrix material containing human skin mimic compounds consisting of a proprietary blend of acids and ammonia and formulated in SPLAT (Specialized Pheromone and Lure Application Technology, a material that allow slow release of odour). The product was produced at ISCA technologies (Riverside, US) and supplied in bubble caps form.

c)
Combination. This is a short-hand term for the combined presentation of Vectrax and Skin Lure, the two-volatile organic-compound blends representing plant attractants and human skin attractants, respectively.

Experiment procedures
Laboratory experiments: All experiments were conducted in the insectary held at 27±1°C, 65±5% RH under a 12:12 h light-dark cycle. Adult females 4-5 days old were released into rectangular 91 x 46 x 30 cm mesh cages, 20 mosquitoes per cage, in which tested attractants choices were offered (Fig. 1). In the rst experiment, the attractiveness of Vectrax and Skin-Lure were compared to each other with blood-fed and unfed females. In the second experiment, each attractant, Vectrax and Skin-Lure, was compared against the combination in separate cages on parous and non-parous females. Parous mosquitoes switch in behaviour from nectar feeding to a combination of host feeding and nectar feeding [19]. Comparison between parous and non-parous female mosquitoes may thus highlight differences that are due to this switch in preference.
Each attractant was offered in a 10

Laboratory experiments
In the laboratory experiment, a total of 292 females were captured in the traps, 157 non-parous and 135 parous out of 480 female mosquitoes that were released during the experiment. Floral odors caught signi cantly less than a combination of oral+skin odor (P 0.001), or skin odor alone (not signi cant for nulliparous females). Nulliparous females slightly preferred skin odor over a combination oral+skin odor. However, choices between nulliparous and parous females did not differ.

Semi-eld experiments
In the mosquito-sphere trial recapture rate of nulliparous mosquitoes was signi cantly lower than that or parous mosquitoes (25.9 and 33.8%, p<0.05). Traps baited with oral odor caught signi cantly more nulliparous mosquitoes then either skin odor or a combination of oral odor with skin odor (Fig 2,  p<0.05,). In contrast, parous mosquitoes were equally captured by the two lures and their combination.

Discussion
Odor-based mosquito control tools slowly nd their way into application, thereby diversifying the toolbox available to local vector control schemes. With much of mosquito life revolving around odors, methods that harness a mosquito sense of smell hold great promise in providing novel tools. A broad array of sensory neurons hardwires mosquito preference and tunes its nose to resources important for survival and reproduction. A mosquito's needs, however, frequently changes between mating, nectar feeding, blood feeding and oviposition, and with that the odors to which it orients. Mosquitoes thus have to 'toggle' between sensory modes, which involves peripheral [43,44] and/or central modulatory factors [44].
Here we evaluated whether a combination of odors from spatio-temporally different origins would synergize attraction, or, alternatively, constitute olfactory nonsense to a mosquito nose and potentially mask attractiveness. We show that combining synthetic mimics of oral and human odor attract nulliparous and parous mosquitoes. As eld populations are comprised of mosquitoes whose odor preferences vary with e.g. age, nutritional and gonotrophic state, such complex, multiplexed blends may be more effective and take a broader sweep of the mosquito population.
Over the past sixty years, attraction of female mosquitoes to blood-host mimicking odours and plantbased attractants has rarely combined odors from different origin [45][46][47][48][49][50][51][52]. In those studies where odors of presumed oral and vertebrate origin were combined, mixed results were obtained, by and large not indicating synergy [37][38][39][40]. In our study, although mosquitoes were attracted by a combination of oral and human odor, the combination did not augment or synergize capture rates, in spite of each blend individually being attractive. This is largely in line with earlier reports [37][38][39][40]. Somehow the added sensory input does not translate in an enhanced 'attractiveness' of the signal. This could in part be because the nutritional status of the mosquitoes in our cohorts was similar, whereas 'synergy' or 'augmentation' of trap capture for a multiplexed lure would more readily emerge in eld populations with mosquitoes in diverse physiological states. Further, it may also be that that odor sources, although placed in very close proximity of each other, do not create fully merged plumes, which mosquitoes may perceive as two separate sources instead of an augmented single source. Indeed, insects are exquisitely capable of neurologically parse incompletely mixed strands of odors [53,54].
Of further interest is the observation that in semi-eld experiments nulliparous females preferred oral volatiles to other blends, whereas this preference disappeared in parous females. This demonstrated a well-known mosquito food proclivity: female mosquitoes generally take sugar meals before they seek a blood meal, and some species strongly prefer sugar over blood or rarely bite until after a sugar meal or even not until after several weeks of sugar feeding [19,55]. The relatively young (4-5 days old) and nulliparous females in this study may thus follow such pattern and rst cater to their low energy reserves before seeking blood. In contrast, parous females, which likely have increased their energy levels through a previous bloodmeal were equally attracted to either lure, as they are known to alternate between sugar meals and blood meals [19].
Combining the oral and skin odor blends is also of interest as they induce sensory activity in entirely different classes of sensory neurons, with oral odors being detected by olfactory receptors (ORs), whereas the detection of the human odor blend, consisting of amines and acids, is entirely restricted to ionotropic receptors (IRs) expressed in grooved peg sensilla [56]. Accordingly, the input from oral and human odor is complementary and induces responses in separate olfactory subcircuits [57]. Combination of input from these subcircuits often lead to synergistic trap catches in other insect taxa example in Drosophila ies [58]. In mosquitoes, however, the relative importance of the OR and IR subcircuitry may differ between distinct behaviors, such as orientation to nectar versus blood host resources [59]. How a combination of input from these classes of sensory neurons in uences capture rates in mosquitoes, and e.g. synergize capture rates of each blend separately, has not been systematically analysed. The tests performed here indicate that different from some other insect taxa, IR and OR input does, perhaps surprisingly, not necessarily synergize. Whereas this may indicate a fundamental odor-coding difference between the taxa, it may also simply be due to that the combination, release rates and ratios require further adjustment.
The results further suggest that a previous blood meal experience modulates olfactory preference. Shifts in blood host preference has been reported for mosquitoes [60,61]. Similarly, shifts in preference have been found depending on internal state, such as age, mating status, physiological status, and blood feeding status [62][63][64]. The modulation observed here, from oral to skin odor, also implies that following a bloodmeal mosquitoes may increasingly 'weigh' input from the IR circuitry, tuned to vertebrate hosts, in behavioural preference. Further research is needed to more in depth evaluate protracted effects of a blood meal on nutritional status and preference modulation, as implied by our results.
From an applied perspective the results offer interesting angles. Although we did not nd any augmentation of trap catches by combining oral and skin odors, the combination did catch both nulliparous and parous mosquitoes (which differs slightly from earlier reports, [37][38][39][40], and would therefore attract mosquitoes relatively independently of physiological status, these being either searching to replenish carbohydrate energy reserves for ight and maintenance ( oral odors) [19,65,66] or searching for hosts to support reproduction (skin lure, a human skin-mimicking blend of volatiles) [28,67,68]. Although there was signi cant attraction to either oral and skin odor blends, alone and in combination, to both parous and non-parous females, this study did not compare the attraction of the blends to a living human. Further research is needed to assess the attractiveness of the new odour blends compared to that of humans in natural eld settings.

Conclusion
Multiplexing volatiles of spatio-temporally segregated odor sources can attract mosquitoes in different physiological state. Captures with such a bait may sample mosquito populations more broadly and represent mosquito populations more accurately. In addition, such lures may be used in novel attract-andkill methods that not only attract young and nulliparous mosquitoes out for a carbohydrate source, but also parous and blood-host seeking individuals that may already be infected with malaria, and thus doubly impact mosquito longevity and malaria transmission. Fine-tuning such lures to target mosquitoes selectively can further increase e cacy, environmental friendliness and prospect in future application.

Declarations
The authors declare that they have no competing interests Ethics approval and consent to participate Availability of data and material: All available data are included in this article.
Competing interests: All authors declare no con ict of interest.
Funding: This study received nancial assistance from Bill and Melinda Gates Foundation through ISCA Technologies, Inc.   Scheme of attractants testing setup in cage trial under laboratory settings