The effects of larval organic fertiliser exposure on the larval development, adult longevity and insecticide tolerance of zoophilic members of the Anopheles gambiae complex (Diptera: Culicidae)

Zoophilic members of the Anopheles gambiae complex are often associated with cattle. As such, it is likely that the immature aquatic stages will be exposed to cattle faeces as a pollutant. This study aimed to examine the effect of cattle manure on members of the An. gambiae complex found in South Africa. In this study, a commercial organic fertiliser originating from cattle manure was used as a proxy for cattle faeces. Laboratory strains of An. merus, An. quadriannulatus as well as four An. arabiensis strains (SENN and MBN: insecticide susceptible, MBN-DDT: insecticide resistant, unselected, SENN-DDT: insecticide resistant: selected for resistance) were used in this study. The effect of larval fertiliser exposure on larval development rate and adult longevity was assessed in all three species. The effect of larval fertiliser exposure on subsequent adult size, insecticide tolerance and detoxification enzyme activity of the four strains of the malaria vector An. arabiensis was also assessed. Following fertiliser treatment, all strains and species showed a significantly increased rate of larval development, with insecticide susceptible strains gaining the greatest advantage. The adult longevities of An. merus, An. quadriannulatus, insecticide susceptible and resistant An. arabiensis were significantly increased following fertiliser treatment. Insecticide susceptible and resistant An. arabiensis adults were significantly larger after larval organic fertiliser exposure. Larval fertiliser exposure also increased insecticide tolerance in adult An. arabiensis, particularly in the insecticide resistant, selected strain. This 4.7 fold increase in deltamethrin tolerance translated to an increase in pyrethroid resistance intensity, which could exert operational effects. In general, larval exposure to cattle faeces significantly affects the life histories of members of the An. gambiae complex.


Introduction
The Anopheles gambiae complex consists of eight morphologically indistinguishable species that differ in their behaviour and malaria vectorial capacity [1,2]. Despite these differences, their larval environments are generally similar. All member species of the complex breed in temporary bodies of water that are usually sunlit, shallow and clean [2,3]. This however appears to be changing as members of the complex, the major malaria vector species An. gambiae and An. arabiensis in particular, will also breed in polluted water [4,5]. This has numerous effects on their life histories and, most importantly, selection for resistance to insecticides [6,7].
The role of pollution as a mode of selection for insecticide resistance is well examined-see review by [8]. When examining pollution, however, there is a tendency to examine anthropogenic pollutants, usually toxic in nature, such as heavy metals or agrochemicals such as herbicides [9,10]. Even more innocuous pollutants such as inorganic fertilizers can also affect life history [11,12]. Very little attention has so far been paid to organic pollutants, despite the close relationship between cattle and the An. gambiae complex, with the relationship of An. arabiensis, as well as other zoophilic members,and cattle particularly important.
An. arabiensis is partially characterised by its behavioural plasticity, and is generally described as exophagic and exophilic [3]. An. arabiensis are zoophilic nocturnal feeders but are also opportunistic feeders, particularly of humans. Importantly, this species is known to adopt both endophagic and endophilic behavioural patterns if potential hosts primarily reside indoors [13]. This behavioural plasticity means that this species is not easily controlled by traditional methods that are based on the indoor deployment of insecticides [14], leading to ongoing residual transmission despite control interventions [15]. An. arabiensis is a primary vector of malaria in South Africa [16], and resistance to both DDT and pyrethroids have been reported in KwaZulu-Natal, one of the country's malarious provinces [17,18]. This species therefore represents a threat to South Africa's malaria elimination agenda.
An. arabiensis tends to flourish in agricultural regions. As such, the relationship between this species and the agricultural and urban industry is an important one [19][20][21]. Additionally, An. arabiensis tends to flourish in association with maize and rice farming [21][22][23]. This species is especially closely associated with cattle as a primary blood source [24,25] and is therefore highly likely to be exposed to cow dung that pollutes breeding sites such as rice paddies [26]. Within the An. gambiae complex, the non-vector An. quadriannulatus and the minor malaria vector An. merus are often found breeding in the same habitats as the major vector An. arabiensis. An. quadriannulatus has never been implicated in malaria transmission and is an endophilic, endophagic and zoophilic species [3]. DDT resistance has been reported in South African An. quadriannulatus [17]. Anopheles merus is a localised malaria vector, and this exophilic, exophagic species is notable for being capable of breeding in sites with a salt content that is usually not tolerable to other mosquitoes [2,3]. In districts of Mpumalanga Province, South Africa, increasing numbers of An. merus and An. quadriannulatus are being detected [27]. The larval dynamics of sympatric species are important, as they could have potential effects on malaria transmission dynamics [28,29] The presence of inorganic pollutants appear to favour the development of insecticide resistance [30], and insecticide tolerant larvae are at an advantage in polluted conditions [31,32], but usually at a biological cost [33]. For inorganic pollutants and various organic agrochemicals this is not surprising, as their relatively late introduction into the environment means that mosquitoes would not have evolved to cope with them as they would have phytochemicals. It must be noted that in some pollutants, notably the herbicide glyphosate, hormetic effects have been noted [34,35]. Hormesis is the phenomenon whereby an exposure compound may confer fitness advantages or disadvantages depending on the dose (reviewed in [34]). Organic water body pollutants such as cattle faeces are not regulated as are other water pollutants, as they are non-toxic.
Despite the close association between agricultural activities, the presence of cattle and the incidence of malaria, very little information is available concerning the role of cattle waste on malaria vector life histories. A single study reported an increase in both Anopheles and Culex larvae breeding in sewage-contaminated water, hypothesising that the faecal matter increased the nutrient content of the water, resulting in increased larval density. High culicine density was accompanied by low anopheline density [36]. Another study noted that increased organic waste increased the efficacy of methoprene for larval control [37]. The aim of this project was therefore to examine the effect of organic pollutants on the life histories of certain member species of the An. gambiae complex, with a special focus on the effect on An. arabiensis.

Materials
All Anopheles mosquito strains used in this project were housed in the Botha de Meillon insectary and reared as described in [38]. In brief, larvae were reared in reverse osmosis water at 25˚C (±2˚C) and 80% relative humidity (±5%) with a 12 hour light/dark cycle with a 30 minute dusk/dawn cycle. Larvae were fed a mixture of powder Beano dog biscuits and yeast.
Anopheles arabiensis. SENN: An insecticide susceptible strain, originating in Sennar, Sudan. It has been in colony since 1980.
SENN-DDT: An insecticide resistant strain selected from SENN, and has been continuously selected for DDT resistance. It is resistant to multiple insecticides, is fixed for the kdr L1014F mutation and has elevated detoxification enzyme levels [39][40][41].
MBN: An insecticide susceptible strain from KwaZulu-Natal Province, South Africa. MBN-DDT: An insecticide resistant strain selected for DDT resistance. This strain is not currently under selection, but still displays resistance [42]. The basis for resistance in this strain is primarily metabolic [43].
Anopheles merus. MAFUS: AN insecticide susceptible strain, established in 2012 and originating from Mpumalanga Province, South Africa. The strain is reared in 8.5% w/v salt water.

Methods
The commercial organic fertiliser (OF) Fertilis was used as a proxy for cattle faeces (Fertilis-Fertiliser made from dairy cow manure processed by earthworms; reg number: B3664 Act 36/1947; Planner Bee Plant care, Kyalamyi, South Africa). An attempt was made to determine a lethal dose, but the fertiliser was found to be non-toxic as no significant mortality could be determined. Therefore, for the sake of water clarity, the dose for all treatments was set at 0.5% (V:V).

The effect of larval OF exposure on development time
For each strain, 100 first instar larvae (less than 24 hours old) were exposed to OF-polluted water. Larvae reared in clean water served as a control. All treatments were fed the same amount of food and were maintained under standard insectary conditions. The time to pupation was monitored. The experiment was replicated three times with larvae originating from three different egg batches per strain.

The effect of larval OF exposure on adult longevity
Samples of larvae from each strain were reared while exposed to OF (except for the controls) as described for the development experiment. Pupae were collected and the adults were allowed to emerge. Thirty males and females of each strain were collected and their longevity was monitored until death. Adults were allowed ad libitum access to a 10% sucrose solution, and the females were not offered blood meals. Cadavers were removed daily. The experiment was replicated in triplicate.

The effect of larval OF exposure on An. arabiensis adult size
Samples of larvae from the insecticide susceptible An. arabiensis strains SENN and MBN, as well as the insecticide resistant SENN-DDT and MBN-DDT were reared in OF-treated water as described for the development experiment. Thirty males and females of each of strain were cold-killed and their wings were removed. Wing length was measured using an Olympus SZX7. Wing lengths were measured from wing tip to allula and used as a proxy for adult size [44]. The sizes of OF-treated adults were compared to those of adults reared in clean water.

The effect of larval OF exposure on An. arabiensis adult insecticide lethal time (LT50)
Samples of larvae from the insecticide susceptible An. arabiensis strains SENN and MBN, as well as the insecticide resistant SENN-DDT and MBN-DDT were reared in OF-treated water as described for the development experiment. Three days old, non-blood fed females and males of equivalent age were exposed to discriminating concentrations of malathion and deltamethrin using CDC bottle bioassays to determine the lethal time to 50% mortality (LT50). The insecticide susceptibility assays are described in [32]. In brief, a dosage of 10μg/ml and 1μg/ml of either malathion or deltamethrin was used to coat bottles for SENN-DDT and SENN respectively. A range of exposure times were used to determine the respective LT50s: 2, 4, 8, 16 and 32 minutes for SENN and 10, 20, 40 and 80 minutes for SENN DDT. The varying exposure times and concentrations were chosen to compensate for the differences in insecticide susceptibility between strains. The experiment was replicated three times with larvae originating from three different egg batches per strain.

The effect of larval OF exposure on An. arabiensis adult insecticide resistance intensity
Due to the effect of larval exposure to OF on subsequent SENN-DDT adult insecticide tolerance, the effect of OF exposure was also examined on pyrethroid resistance intensity [45]. SENN DDT first instar larvae were reared in OF-polluted water, with larvae reared in clean water as a control. Pyrethroid resistance intensity was determined as described in [42]. The experiment was replicated three times with larvae originating from three different egg batches per strain.

The effect of larval OF exposure on An. arabiensis adult detoxification enzyme activity
SENN, SENN-DDT, MBN and MBN-DDT larvae were reared in OF-treated water, with control larvae reared in clean water. All treatments were fed the same amount of larval food. Adults were subsequently harvested and 48 males and females of from all treatments were cold-killed at the age of three days after access to sucrose only. These adults were then homogenised in PCR-grade water. Cytochrome P450 activity, measured as haeme peroxidase activity, Glutathione S-transferase activity, general esterase activity and catalase activity was assessed calorimetrically as described in [31].

Statistical analysis
Data were analysed using Statistix 8 (Analytical Software, Tallahassee, Florida) and IBM SPSS statistics version 22 (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY). Data were analysed for normality using the Shapiro-Wilk test. Differences in normally-distributed data were analysed using 1-way Analysis of Variance or 2-sample ttest. Non-parametric data were analysed using a Kruskal-Wallis one-way non-parametric analysis of variance. Longevity was assessed using the Kaplan-Meier estimator, with the Log Rank test used as a measure of significance. Data were analysed at a confidence interval of 95%.

The effect of larval OF exposure on adult longevity
Adults of the insecticide susceptible strain SENN lived significantly longer when the larvae were exposed to OF (Log rank test: p = 0.04, χ 2 = 8.32, DF = 4) (Fig 2A). This was also true for the susceptible MBN strain (Log rank test: p<0.01, χ 2 = 16.26, DF = 4) ( Fig 2B). There was no significant difference in adult longevity in SENN-DDT after larval OF exposure (Fig 2C). The MBN-DDT strain also showed a significant increase in adult longevity after larval OF exposure (Log rank test: p = 0.02, χ 2 = 11.32, DF = 4)( Fig 2D).

The effect of larval OF exposure on An. arabiensis adult insecticide lethal time
A similar pattern was observed for malathion resistance in SENN-DDT and MBN-DDT females, where larval OF treatment resulted in a significant increase in malathion LT50 (1-way  (Fig 4B).

The effect of larval OF exposure on An. arabiensis adult insecticide resistance intensity
SENN DDT adults had the longest pyrethroid lethal time, and were therefore used to determine whether this increase in tolerance resulted in an increase in pyrethroid resistance intensity. Untreated SENN DDT showed a moderate resistance intensity to permethrin, but this was increased to high after larval OF treatment in both males and females. Similarly, SENN DDT had a low resistance intensity to deltamethrin (as defined in [42]), but after OF treatment resistance intensity increased to high (Table 1).    activity in males and females (Kruskal-Wallis One-way AOV: females: p = 0.02, F = 6.17, DF = 1; males: p = 0.03, F = 4.77, DF = 1) (Fig 5B).

Consumption of organic fertiliser by larvae
To determine whether any of the effects of observed were due to the consumption of the organic fertiliser, fourth instar larvae were examined for evidence of the material in their gut. Fig 6A and 6B represent a larvae reared in clean water and its' resected midgut. There was no evidence of fertiliser in the food bolus. Fig 6C and 6D represents a larvae reared in organic fertiliser and its' midgut respectively. There is evidence of organic fertiliser in the food bolus of the resected gut. Fig 6E shows a comparison of a treated and untreated 4 th instar larvae, and Fig 6F show the content of those larvae's midguts, demonstrating the presence of organic fertiliser in the gut, suggesting that the larvae are ingesting the fertiliser.

Discussion
The advantage conferred by larval exposure to organic fertiliser appears to be linked to the consumption of fertiliser by the larvae (Fig 6). This is suggested by the presence of the fertiliser in their guts. It has been reported that larval nutrition positively affects numerous life history traits, particularly in An. arabiensis [23,39]. Although the effect of the consumption of organic detritus has been examined in container breeding mosquitoes such as Aedes albopictus, Ae. aegytpi and Culex quinquefasciatus [46][47][48], the effect is less clear in Anopheles mosquitoes. The aforementioned studies all examined organic detritus with particular reference to plant matter such as leaf litter, which would have a greater effect on Aedes mosquitoes than

Anopheles.
A previous study demonstrated that the presence of cattle faeces enhanced the Anopheline population in a rice growing area under semi-field conditions [26]. The basis of this observation still remains unexplored. This study partially addresses this question. What is of interest about this exposure is that unlike exposure to inorganic or more toxic pollutants, insecticide susceptible as well as resistant mosquitoes gain an advantage. The advantages to insecticide susceptible strains are numerous, including shorter larval development time and increased adult longevity. Importantly, these observations were conserved across the member species of the An. gambiae complex assessed here. Insecticide resistant An. arabiensis seem to gain an advantage primarily in terms of increased insecticide tolerance. The effect of continued insecticide selection pressure appears to play a crucial role in this effect. While larval exposure also increases larval development rate, the size and longevity advantages seen in susceptible strains were not evident in the insecticide resistant, selected SENN-DDT strain. Although MBN-DDT was selected from MBN it retains its' resistant phenotype despite the fact that selection has ceased. Therefore, it is representative of a strain that is no longer under selection pressure. This lack of difference in developmental rate between the two strains suggests a lack of fitness cost in the MBN-DDT strain. The increased rate of development in MBN mirrors the findings in SENN and SENN-DDT. This highlights the finding that insecticide susceptible strains gain a greater advantage in terms of increased development rate than insecticide resistant strains.
The marked difference between the advantages gained by insecticide resistant and susceptible strains is also observed in longevity. Like with larval development, insecticide susceptible strains gained a marked advantage in increased longevity. This may be due to nutritional advantages gained by the consumption of the organic fertiliser, and that the advantages in the insecticide susceptible stains are diffuse, resulting in increased rate of development, size (an indicator of fecundity [49,50]) and longevity in insecticide susceptible strains. It is notable that a sex-specific difference in changes in longevity occurs in An. merus. Although there is no confirmation about why this may be so, it may be due to sex-specific preferences in the consumption of the organic fertiliser.
A previous study on the effect of nutrition of An. arabiensis insecticide resistance demonstrated that larval nutritional deprivation had a greater effect on the insecticide susceptible SENN strain than the resistant SENN-DDT strain [39]. These data, along with the findings of this study, suggested that the greatest advantage for the insecticide resistant, selected SENN-DDT strain was increased insecticide tolerance. This suggests that in this strain resource allocation is directed towards maintenance of the insecticide resistance phenotype, as metabolic resistance is an energetically costly process [51,52].
The differences by which the MBN-DDT and SENN-DDT mediate increased insecticide tolerance are also worth noting. While the process appears to be mediated by increased cytochrome P450 activity in the MBN-DDT strain, general esterase activity was most markedly increased in the SENN-DDT strain. OF treatment significantly increased alpha-esterase activity in males, which may underlie the increase in male lethal time for both malathion and deltamethrin. The decreased deltamethrin lethal time is not explained by detoxification enzyme activity. Furthermore, the decrease in GST activity, coupled with a decreased catalase activity in the SENN-DDT strain, suggests a lowered requirement for oxidative stress defence, in turn suggesting a lowered oxidative stress burden in this strain, which is associated with increased insecticide resistance [53,54]. This, however, would require measurement of oxidative stress markers.
Another mechanism whereby organic fertiliser pollution may be affecting insecticide resistant phenotypes is by the modulation of gut microflora. Larvae acquire their gut microbiome from the aquatic environment [55] and, therefore, the consumption of organic fertiliser may alter the gut bacterial composition. Gut bacterial composition has been demonstrated to play a crucial role in both life history and expression of insecticide resistance phenotypes [56].
The SENN and SENN-DDT strains have recently been used to test the effects of a number of stressors on the expression of insecticide resistance. It is worth noting that larval organic fertiliser caused the highest recorded increase in lethal time in SENN-DDT, more than herbicide or metal exposure [31,32]. Crucially, this resulted in an increase in insecticide resistance intensity. Insecticide resistance intensity is one of the most crucial markers of the operational impact on vector control [42]. Therefore, the increase in deltamethrin and permethrin resistance intensity from low and moderate to high suggests that exposure to organic fertilizer at the larval stage can assist in producing resistance phenotypes at the adult stage that may have be operationally significant [35].
In conclusion, members of the An. gambiae complex are exposed to cattle manure during their immature aquatic stages. Using organic fertiliser as a proxy for cattle faeces, it is evident that mosquito larvae feed on the fertiliser which then confers numerous life history advantages (reduced larval development time, increase adult longevity, increased adult size, and increased insecticide tolerance and intensity), particularly for insecticide susceptible strains. The effect on increasing insecticide tolerance is most marked in insecticide resistant strains. The marked increase in insecticide tolerance and translation into an increase in pyrethroid resistance intensity in association with OF treatment, is predicted to produce a resistant phenotype that could exert an operational impact.
Supporting information S1