Plant Nutrient Quality Promotes Survival and Reproductive Fitness of the Dengue Vector Aedes Aegypti

Background: In a recent study using DNA barcoding, we identied the plants fed upon by four Afro-tropical mosquito species that vector dengue, malaria, and Rift Valley fever. Herein, we have expanded on this study by investigating the role of three of the plants Pithecellobium dulce (Fabaceae), Leonotis nepetifolia (Lamiaceae), and Opuntia cus-indica (Cactaceae) on survival, fecundity, and egg viability of the dengue vector Aedes aegypti. Methods: We tested these effects using females that received a) an initial three rations of bloodmeal, and b) received no bloodmeal at all. Two controls were included; age-matched females fed on glucose solution with or without initial bloodmeal, and those fed exclusively on bloodmeal. Data was collected daily over a 30-day period. The amino acid content of Ae. aegypti guts and the amino acid content of their respective diets was detected by coupled liquid chromatography – mass spectrometry. Results: Females fed on P. dulce and exclusive bloodmeal had a shorter survival than those fed on glucose. On the other hand, females fed on L. nepetifolia survived longer than those fed exclusively on bloodmeal, whereas those fed on O. cus-indica had the shortest survival time. With initial bloodmeal, females fed on L. nepetifolia laid 1.6-fold more eggs while those fed on the other diets laid fewer eggs, compared to those fed exclusively on bloodmeal. Hatching rates of the eggs laid varied with the diet. Mass spectroscopic analysis of gut contents of mosquitoes exposed to the different diets showed qualitative and quantitative differences in their amino acid levels. Conclusion: Our ndings highlight the central role of plant in the reproductive tness of dengue which may impact their disease


Background
The last two decades has seen the resurgence and spread of arboviruses such as dengue, zika and chikungunya viruses that are vectored by Aedes mosquitoes. Although the number of dengue cases are underreported and misclassi ed [1], the number of global infections is estimated at 390 million annually, with about 50-100 million cases manifesting clinically [2,3]. The geographic expansion of dengue, which is caused by four dengue virus serotypes (DENV 1-4), has been characterized by increase in case incidence, epidemics and super-endemicity, with more frequent severe forms of dengue [4,5]. The recent outbreak of Zika in the South and Central America, and the Caribbean region attests further to the continued geographic spread of arboviral diseases [6,7]. Zika was rst identi ed in Rhesus monkey in Uganda in 1947, with the rst human cases detected in 1952 in Uganda and Tanzania [8]. The Zika outbreaks in Brazil and Colombia in the year 2015, and its subsequent spread to 13 other countries in the Americas, along with other outbreaks in the Paci c (Yap, 2007;French Polynesia, 2013) and Africa (Cape Verde, 2015), highlights the growing concern of the rapid expansion of arboviral diseases [9]. The overarching commonality among these diseases is that there are no speci c drugs currently for their treatment and no viable vaccines available [10]. This makes effective vector control the mainstay for prevention and control of these diseases.
The geographic expansion of these diseases closely follows the tropical and subtropical distribution of their primary vectors, Aedes aegypti and Aedes albopictus. The spread has been attributed to a range of factors including climate change, uncontrolled urbanization, globalization, travel, trade, socioeconomics and the ability of these viruses to evolve [5,11]. In addition, factors contributing to the resilience of Ae. aegypti populations such as insecticide resistance, the ability of eggs to withstand desiccation or undergo diapause, ability of adults to adapt to environmental modi cations and their behavioral plasticity have contributed to the sustenance or even expansion of Ae. aegypti populations [12,13]. Overall, the continued expansion of the geographic range of these mosquito species and the pathogens that they transmit calls for a detailed understanding of vector and disease ecologies in the renewed effort for innovative management strategies.
Plant feeding is emerging as a key ecological factor in the biology of several mosquito species including Aedes species [14][15][16][17]. While plant feeding pre-dates blood feeding in insects, blood sucking arthropods are thought to have adopted the latter trait during evolution to enhance the propagation of their progeny [18]. Among different mosquito species, intermittent plant feeding in females has long been documented but its role with respect to reproductive tness has been downplayed by different studies [19][20][21][22]. Central to this dogma are Ae. aegypti and Anopheles gambiae, the two highly anthropophilic and most important disease vectors. Variably low fructose levels detected in eld collected females of these species accompanied by their tendency to have multiple blood meals has led to the proposition that they seldom feed on plants but depend on human blood for both their metabolic processes and reproduction.
However, recent evidence where more sensitive trapping strategies and analytical approaches were used shows of higher plant feeding frequencies in these two species [16,17,23].
Several studies have demonstrated the central role played by plant sugars in male and female mosquito survival, mating competence and ight activity. In addition, there has been substantial effort to identify plant species fed upon by different mosquito species. These efforts have greatly been boosted by the advent of highly sensitive analytical techniques such as plant DNA barcoding and mass spectrometry which provide secure host plant identi cation and authenticate their trophic association [24,25].
Molecular approaches have recently been used to identify plant species fed upon by important disease vectors such as An. sergentii [23], An. gambiae, Ae. mcintoshi, Ae. ochraceus and Ae. aegypti [16], and phlebotomine sand ies [26,27] in their natural habitats. Evidence of more frequent plant feeding among these vectors and the identi cation of host plant species further augment the proposition of their central role in vector population dynamics. However, beyond a few studies linking plant feeding to mosquito survival, little is known about the nutritional contribution of plants to vector tness and population dynamics.
Building on our recent identi cation of natural host plants of four Afro-tropical mosquito species [16], we sought to elucidate the role of plant nutrition using three of the identi ed plants on survival and reproductive tness of Ae. aegypti.

Methods
Experimental animals F1 generation obtained from Ae. aegypti eggs collected in Kili (3.6333° S and 39.8500° E) in the coastal region of Kenya endemic for dengue (Sang and Dunster, 2001) were used. The eggs were collected by placing black ovicups lined with brown ovistrips in pre-identi ed A. aegypti breeding sites overnight. The collected eggs were either hatched immediately or carefully dried and transported to icipe laboratories in Nairobi. The hatching larvae were reared in plastic trays (25 cm long × 20 cm wide × 14 cm high) to adults with a daily ration of Tetramin sh food (Tetramin1, Melle, Germany) of 0.3 g /100 larvae/day. The rearing room was maintained at a temperature of 28 (± 1) o C and relative humidity of 80 (± 5) % and a photoperiod of 12: 12 (light: dark) hours. The haplotype of the emerging adults was all con rmed to be Ae. aegypti aegypti (hereafter referred to as Ae. aegypti) as they all had white scales on the rst abdominal tergite (McClelland,1960). The adults, 1-2 days old with no prior exposure to any other nutrient source, were used in survival and fecundity assays.

Plant materials
Plant species identi ed as natural host plants of the four mosquito species from our previous study [16] were used in these assays although speci c for the vectors from their respective ecologies. These transported to icipe laboratories in Nairobi. They were used when they started to blossom.
All the experiments were conducted under controlled conditions as described above for mosquito rearing.
Survival, fecundity, and egg hatchability of Aedes aegypti on different host plants In Experiment I carried out in Kili , two assays were conducted. In the rst assay, a group of 100 males and 100 females were introduced into a 30 × 30 × 100 cm cage containing P. dulce cuttings. In addition to the plant, which was continuously available, they were provided with initial three mice blood meals at day 3, 5 and 7 from the onset of the assay. The blood meal was provided by placing anaesthetized mouse on top of the mosquito cages and the mosquitoes allowed to feed on them for an hour. Oviposition cups were provided in all the cages 48 h after the rst bloodmeal. They were monitored for survival and fecundity daily for 30 days. Mortality and the daily number of eggs laid were recorded. Control experiments comprised 100 female and 100 males Ae. aegypti with access to a) 6% glucose solution plus three initial blood meals and b) blood meals only on alternating days for thirty days with a total of 15 blood meals. A total of three replicates using three different batches of mosquitoes were carried out for all nutrient regimes. Nine living female mosquitoes were randomly selected from each replicate of all the treatments on day 15 (chosen to avoid the confounding effect of blood derived amino acids in the gut of mosquitoes fed on plant diet) for amino acid analysis as described below. The second assay was the same as the one above, but no blood meal was provided.
Similar experimental set up as above was used in Experiment II for survival and fecundity assays using L. nepetifolia and O. cus-indica. Newly emerged females and male (100 mosquitoes for each sex) were provided with either L. nepetifolia, O. cus-indica, 6% glucose solution or mice bloodmeal provided on alternate days. Except for the latter group, the mosquitoes in all the other three groups were either provided an initial three mice bloodmeals on days 3, 5, and 7, or no bloodmeal at all. Survival, fecundity, and mosquito sampling for amino acid analysis was done as described above.
To measure the hatchability of the laid eggs, the eggs were put in 18 × 12.5 × 2.1-inch trays and distilled water added to a depth of one inch. The eggs were hatched according to the date laid and nutrient source. The number of larvae were counted daily for up to two weeks after which the unhatched eggs were considered not viable. The counted larvae were promptly removed.

Analysis of host plant amino acid content and the ingested equivalence in Aedes aegypti
To understand the differences in the performance of Ae. aegypti on different nutrient sources, we quanti ed the amount of amino acids in the three plant species and the corresponding amounts ingested by the mosquitoes. Both plant sap from phloem in the succulent tissues and nectar from the nectaries were collected from the three plants using 20 µL micro-capillary tubes (Drummond Scienti c Company, Brumall, PA, USA) tapered on one end using glass puller. Between 10-30 µL of plant sap and nectar were separately collected, transferred into 1.5 mL low-binding eppendorf tubes, snap frozen in liquid nitrogen and immediately stored at -80 o C until analysis. Up to 100 µL of venous blood was drawn from mouse facial vein into 1.5 mL low-binding eppendorf tubes and immediately snap-frozen in liquid nitrogen and stored at -80 o C. Blood was collected from three mice drawn from different litters for both experiments conducted at KEMRI-Wellcome Trust laboratories and icipe. The mosquito samples were prepared by dissecting mosquitoes preserved from survival assays and pooling the gut plus crop (hereafter referred to as gut) from three mosquitoes.
To detect amino acid content, the pooled guts, 10 µL of plant sap + nectar or 10 µL of blood samples were The amino acids were identi ed by comparing their mass spectra with literature data [28].

Statistical analyses
The difference in survival times of adult Ae. aegypti on different nutrient sources was detected using Kaplan-Meier and Cox regression survival analyses. Mosquitoes sampled for nutrient analyses and those surviving after the 30-days observation period were treated as censored. Differences in fecundity between mosquitoes held on different nutrient sources were detected using zero-in ated GLM. The differences in hatching rate from different nutrient sources were compared using one-way ANOVA and Tukey Post Hoc test. The gut amino acid content was quanti ed for the different nutrient sources and the differences detected using one-way analysis of variance. All statistical analyses were done in R software version 3.6.3 [29].

Results
Host plants variably support Ae. aegypti survival, fecundity, and egg viability In Experiment I, survival of Ae. aegypti females provided with an initial bloodmeal and fed on the different diets was signi cantly different (Log rank = 40.785, df = 2, p < 0.001; Fig. 1A); those fed on glucose solution, P. dulce and blood having mean survival of 23.5 ± 0.6, 17.7 ± 0.7 and 16.1 ± 0.7 days, respectively. With no initial bloodmeal, the mean survival of female Ae. aegypti on glucose and P. dulce were 23.6 ± 0.8 and 13.1 ± 0.8 days, respectively (Log rank = 48.04, df = 1, p < 0.001; Fig. 1B). Similar survival patterns were observed in males, with those fed on glucose having a mean survival of 21.6 ± 0.7 days while those fed on P. dulce had a median survival of 14.6 ± 0.7 days (Log rank = 25.162, df = 1, p < 0.001; Fig. 1C).
For fecundity, females fed on P. dulce and glucose laid 1.6-and 2.2-fold less eggs, respectively, than those fed exclusively on blood but no signi cant difference was detected (F (2, 267) = 1.985, I = 0.139; Fig. 2A). On the other hand, those fed on L. nepetifolia laid 1.6-fold more eggs than those fed exclusively on blood, while mosquitoes fed on O. cus-indica and glucose had 1.7-and 2-fold less eggs than those fed exclusively on bloodmeal, respectively (F (3, 356) = 3.495, p = 0.0158, Fig. 2B). Besides having the highest fecundity rate, mosquitoes fed on L. nepetifolia had a sustained moderate oviposition throughout the experimental period which was comparable to those that exclusively fed on bloodmeal ( Fig. 2C and D).
Variable amino acid quality support observed differences in the tness matrix of Ae. aegypti fed on different host plants A total of 12 amino acids present in mice blood were detected in variable amounts in the sap plus nectar of three plant species. These included valine, serine, glutamine, proline, glycine, methionine, tyrosine, isoleucine, leucine, phenylalanine, tryptophan, and arginine. Uniquely abundant amino acids detected in the guts of mosquitoes fed on the nutrient regimes included valine, arginine, isoleucine, methionine, and phenylalanine (Fig. 4A). Valine was 5-, 13-and 14-fold more abundant in the guts of mosquitoes fed on L. nepetifolia, glucose and O. cus-indica, respectively. In addition, arginine was abundant in the guts of those fed on O. cus-indica, isoleucine in those fed on P. dulce, and methionine in those fed on L. nepetifolia (Fig. 4B). On the other hand, phenylalanine was 6-, 8-and 11-fold less abundant in the guts of mosquitoes fed on P. dulce, O. cus-indica and glucose solution, respectively, relative to those exclusively fed on blood (Fig. 4B). Notably, glutamic acid was present in the guts of mosquitoes fed on mice blood but absent in those from all the other diets. To further con rm that female Ae. aegypti indeed were able to imbibe these amino acids from their host plants, we analyzed for four of the identi ed amino acids in the guts of non-blood fed mosquitoes. Besides valine, none of the females from glucose diet had any detectable amino acids in their guts. However, females fed on all the three host plants had variable amounts of methionine, isoleucine, phenylalanine, and arginine in their guts (Fig. 4C).

Discussion
Our ndings show that the three plants used in this study differentially impact on the survival and reproductive tness of dengue vector, Ae. aegypti. We previously reported a sugar feeding frequency of 17% in female Ae. aegypti collected around vegetations in the coastal Kenya [16]. The study by Olson et al. [17] demonstrated that sugar feeding occurs at a much higher frequency than previously reported, with collection method and season being important in in uencing the proportion of fructose-positive females captured. Plant sugars, particularly fructose, have been shown to provide a ready source of energy for various metabolic processes in several mosquito species [17,[30][31][32][33]. Extended survival time is pivotal in the transmission of vector-borne diseases as it guarantees completion of extrinsic incubation of the causative agents and increases the chances of multiple infective vertebrate host bites [34]. These ndings further reinforce the argument of the central role played by plants in the biology of Ae. aegypti, contrary to previous beliefs.
While no eggs were laid by mosquitoes exclusively fed on the three plant species, signi cant difference in fecundity were observed when mosquitoes fed on them with initial bloodmeal rations. Those fed L. nepetifolia had slightly higher oviposition than their exclusively blood-fed counterparts, while those fed on P. dulce, O. cus-indica and glucose laid fewer eggs than blood fed females. Similar impacts of plant diets on mosquito fecundity have been observed for An. gambiae [30,35] and Culex pipiens [33]. Although sugar feeding has long been suggested to impact fecundity of Aedes mosquitoes [36,37], to the best of our knowledge, this is the rst evidence directly linking plant feeding to Ae. aegypti fecundity. Plant nectars have been shown to increase mating competence in males of different mosquito species [31,32,38,39]. Sugar has also been shown to be important in inducing egg development in autogenous Ae. albopictus and Cx. pipiens f. molestus [40,41]. The failure of Ae. aegypti to lay eggs without initial bloodmeal in this study is not surprising, although varying degrees of autogeny has been reported among these species in East Africa [42,43]. However, the potential of L. nepetifolia to not only boost their overall fecundity but also induce a sustained oviposition long after the last bloodmeal is noteworthy. Although lower than the hatching rates in eggs from exclusively blood-fed females, a 34% of eggs laid by females fed on L. nepetifolia were viable. The difference in fecundity of mosquitoes held on different nutrient sources observed in this study can be explained under three propositions: 1) males fed on L. nepetifolia had sugar-rich diet and therefore increased mating competence and reproductive output in females, compared to males from the exclusive bloodmeal diet which died off within three days; 2) females held on L. nepetifolia imbibed su cient sugar meals/sap from the succulent plant tissues thereby resulting in constant distention of the abdomen and inducing oocyte maturation following initial bloodmeal, as has been reported in the case of Ae. albopictus [40]; and 3) mosquitoes feeding on the three plants, especially on L. nepetifolia, imbibed not only sugar but also amino acids which supplemented those received from the initial bloodmeal in further boosting their reproduction.
We explored the third proposition further by analyzing amino acid content of female Ae. aegypti held on these plant species and comparing the outputs with those from mosquitoes fed exclusively on blood and their plant sources. A total of 12 amino acids present in mice blood were positively identi ed in the guts of both blood-fed females held on different diets in varying proportions. Notably, mosquitoes held on L. nepetifolia had high methionine content in their guts, while those held on P. dulce had high isoleucine content. Intriguingly, phenylalanine was signi cantly low in females held on P. dulce, O. cus-indica and glucose solution, with mosquitoes fed on the three plant species and glucose lacking glutamic acid. Nonblood fed females held on the three plant species had similar amino acid pro les as those offered initial bloodmeal. These observations support our proposition that female Ae. aegypti imbibed variable amounts of amino acids from these plant species which differentially impacted their fecundity. This was further supported by the detection of these amino acids in the respective host plants the mosquitoes were held on. Different amino acids have been shown to impact differently to mosquito fecundity. Phenylalanine and tyrosine have been shown to be important for the development and tanning of An. gambiae and Ae. aegypti eggs [44,45], while isoleucine is important in follicular maturation and preventing egg resorption in Ae. aegypti [46]. Methionine and leucine have been shown to increase fecundity of green pea aphid, Cyrthosiphon pisum, by enhancing target of rapamycin (TOR) signalling pathway [47]. This study represents the rst empirical evidence of the possible involvement of plantderived amino acids in the reproductive tness of Ae. aegypti.

Conclusion
We conclude that these ndings offer signi cant insight into the role of plant-derived nutrients in the biology and population dynamics of Ae. aegypti in the face of rapidly changing vector and disease ecology driven by climate change and human activities. However, a holistic investigation of the ecological drivers of the spread of arboviral disease vectors and their intrinsic interactions with plants in their ecosystem will be important in understanding their epidemiology and application in control strategies based on plant metabolites. We also appreciate that the use of plant cuttings in the assays with P. dulce, and whole potted plants in the case of L. nepetifolia and O. cus-indica, could have contributed to the observed differences in the performance of Ae. aegypti. This warrant further investigation in standardized assays.

Declarations
Ethical approval and Consent to participate Mice (BALB/c strain) used for mosquito blood feeding in these experiments were supplied by icipe Animal Rearing and Containment Unit. All blood feeding experiments were conducted according to IACUC approved protocols. Approval for the study sought from Kenya Medical Research Institute Scienti c and Ethics Review Unit (KEMRI-SERU) (Project Number SERU 2787).
Consent to participate was not required. The impact of plant nutrients on survival male and female Ae. aegypti. A) Survival curves of female Ae. aegypti on P. dulce, 6% glucose solution and exclusive mice blood, with initial 3 bloodmeal rations offered to those held on P. dulce and glucose solution (P-value < 0.001). (B-C) Survival curves of female and male Ae.aegypti, respectively, fed on P. dulce and glucose without initial bloodmeal (P-value < 0.001). D) Survival curves of females held on L. nepetifolia, O. cus-indica, 6% glucose solution and exclusive mice blood, with the former three offered three initial bloodmeal rations (P-value < 0.001). (E-F) Survival curves of female and male Ae. aegypti, respectively, fed on L. nepetifolia, O. cus-indica or glucose solution without initial bloodmeal (P-value < 0.001). Survival curves denoted with the same different letters are signi cantly different. Differences in survival curves was detected using Kaplan Meyer analysis and Cox regression analyses.   The P -values were < 0.001, <0.01, <0.05 and <0.01 for methionine, isoleucine, phenylalanine, and arginine, respectively. C) Mean amounts of the unique amino acids in the guts of females fed on the ve nutrient sources with no initial bloodmeals. The P -values were < 0.05, =0.25, =0.052 and <0.01 for methionine, isoleucine, phenylalanine, and arginine, respectively. The differences in gut amino acid content was detected by one-way ANOVA and Tukey Post Hoc Test. Bars denoted with different letters are signi cantly different.