Variation in the susceptibility of Anopheles gambiae to botanicals across a metropolitan region of Nigeria

Pesticide resistance is normally associated with genetic changes, resulting in varied responses to insecticides between different populations. There is little evidence of resistance to plant allelochemicals; it is likely that their efficacy varies between genetically diverse populations, which may lead to the development of resistance in the future. This study evaluated the response of Anopheles gambiae (larvae and adults) from spatially different populations to acetone extracts of two botanicals, Piper guineense and Eugenia aromatica. Mosquito samples from 10 locations within Akure metropolis in Southwest Nigeria were tested for variation in susceptibility to the toxic effect of botanical extracts. The spatial distribution of the tolerance magnitude (T.M.) of the mosquito populations to the botanicals was also mapped. The populations of An. gambiae manifested significant differences in their level of tolerance to the botanicals. The centre of the metropolis was the hot spot of tolerance to the botanicals. There was a significant positive correlation between the adulticidal activities of both botanicals and initial knockdown. Hence, knockdown by these botanicals could be a predictor of their subsequent mortality. In revealing variation in response to botanical pesticides, our work has demonstrated that any future use of botanicals as alternative environmentally friendly vector control chemicals needs to be closely monitored to ensure that resistance does not develop.


Introduction
Anopheles mosquitoes are medically important because they are vectors of human diseases such as malaria, filariasis and arboviruses [1][2]. Due to a dearth of effective vaccines and drugs to control these diseases, attention has shifted towards the control of the mosquito disease vector [3], with a reliance on the use of synthetic insecticides [4]. However, concerns such PLOS ONE | https://doi.org/10.1371/journal.pone.0210440 January 9, 2019 1 / 18 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 as environmental toxicity, mammalian toxicity, effect on non-target species and insecticide resistance have long been associated with the use of synthetic insecticides [5][6]. An increase in the development of insecticide resistance in mosquitoes is of significant concern in all vector control programmes [7][8][9][10]. As alternatives, phytochemicals obtained from plants with proven insecticidal efficacy have been tested against mosquito vectors such as Anopheles gambiae [11][12]. Among the arrays of plant materials that have been reported to be effective in controlling this vector are extracts from the seeds of Piper guineense (Piperaceae) [12] and flower buds of clove Eugenia aromatica (Myrtaceae) [13]. Piper guineense is a West African spice plant, commonly called black pepper or Ashanti pepper that contains a cocktail of chemicals, including alkaloids and piperidene [14]. Piper guineense contains naturally-occurring piperine-type alkaloids [15] that have insecticidal properties against Aedes aegypti, Culex quinquefasciatus and An. gambiae [16][17][18]. Ethanolic extracts of P. guineense have also been used for mosquito control [13]. The active ingredient in clove is eugenol, a phenylpropene, which is known to exhibit biocidal properties toward Sitophilus zeamais [19], Dinoderus bifloveatus [20] and Ixodes ricinus [21]. In addition, the ethanolic extracts of E. aromatica can kill mosquito larvae within 24 h [22]. This suggests that applications of clove extracts could work rapidly during outbreaks and epidemics where immediate action is required.
Botanical pesticides have long been touted as attractive alternatives to synthetic chemical pesticides for mosquito control [23][24][25]. The possible reason for this is the development of resistance to virtually all classes of insecticides used for mosquito control [26][27]. However, it should not be assumed that naturally derived pesticides are less susceptible to resistance since they may well have similar modes of action. Thus far, history has shown that the overzealous use of synthetic pesticides has resulted in numerous problems that were unforeseen at the time of their introduction. Hence, tolerance surveys are useful to better understand insecticide tolerance patterns, explain control failures, and have a scientific basis for selection, rotation, and discontinuation of particular insecticides [28].
If we understand how the environment plays a role in the response of Anopheles mosquitoes to botanicals, we may be able to predict their tolerance to botanicals, and by extension, assist in the sustainable control of the vector. An alternative to synthetic insecticides is urgently required as current control of Anopheles mosquitoes in Southwest Nigeria using pyrethroid insecticides is being threatened by developing resistance [29][30]. Substantial evidence points to pyrethroid resistance [31][32] resulting from target-site mutations [33][34] and enhanced insecticide detoxification [35][36] with environmental factors influencing the mosquito responses to pyrethroids [33][34][35][36][37][38]. This study investigated the tolerance of An. gambiae within Akure metropolis in Southwest Nigeria to P. guineense and E. aromatica with a view to revealing probable locational differences and the likely hot spots of potential failure (if any) of botanical insecticide efficacy within the metropolis.

Study area
Akure (Fig 1) is the capital city of Ondo State in the South-Western region of Nigeria which is located at latitude 7.2571˚N and longitude 5.2058˚E of the equator, and situated at a mean elevation of 353 metres above sea level [39]. The climatic condition of Akure is influenced mainly by the rain-bearing southwest monsoon winds from the ocean and the dry northwest winds from the Sahara Desert. High temperatures and high humidity also characterize the climate [39].

Botanicals
The seeds of P. guineense (Fig 2A) and dried flower buds of E. aromatica ( Fig 2B) were purchased from local herb sellers at Owena market (7.1965684N 5.0186405E) in Osun State, Nigeria. The plant materials were pulverized using a Marlex grinder (Model Excella 2431a, Marlex PVT LTD Mumbai, India). Their powders were sieved through a mesh size of 1mm 2 and stored separately in tight lid containers at 28 ± 3˚C and 75 ± 5% RH (relative humidity). Acetone extract of the botanicals was obtained separately using a cold extraction method [40]. This was done by soaking 300g of the powder in an extraction bottle containing 900ml of acetone for 72 hours. Filtration was then carried out using a double layer of Whatman No. 1 filter paper. The extraction solvent was evaporated using a rotary evaporator set at 35˚C to 43˚C with a rotary speed of 138 to 148 rpm for 3-4 hours. The resulting extracts were kept in bottles with tight lids and preserved in the refrigerator.

Collection and rearing of mosquitoes
Anopheles gambiae larvae were collected from 10 locations across Akure metropolis (Fig 3). The containers bearing mosquito larvae were transferred to the Entomology Laboratory of the Biology Department, Federal University of Technology, Akure, Nigeria. Larval identification was carried out under Olympus dissecting microscope (X20) using morphological keys [41,42]. Once the larvae had pupated, they were transferred to a screened cage with dimension 20 x 20 x 20cm for adult emergence. The larvae were fed with yeast and reared at 28 ± 3˚C and 75 ± 5% RH. The adult insects were used immediately after emergence.
Ethics statement. No specific permission was required for the collection activities. This is because the larvae collected, in each location, did not involve endangered or protected species. Also, mosquito larvae were collected from water bodies found within peri-domestic human surroundings and not from protected areas or private lands.

Bioassay
Effect of plant extract on mosquito larvae. Larvicidal activities of the plant extracts were carried out according to a modified WHO standard procedure [43]. Fourth instar larvae of An. gambiae were used for this assay. After preliminary bioassays, 0.5ml of each concentration of P. guineense extract in acetone (0.1%, 0.3%, 0.4%, 0.5% and 0.7% delivering 0.0025μl/ml, 0.0075μl/ml, 0.0100μl/ml, 0.0125μl/ml and 0.0175μl/ml, respectively) was added to 200ml water in 250ml glass beaker. Also, 0.5ml of each concentration of E. aromatica extract in the solvent (2%, 3%, 4%, 5% and 6% delivering 0.050μl/ml, 0.075μl/ml, 0.100μl/ml, 0.125μl/ml and 0.150μl/ml, respectively) was used for the bioassay. Twenty mosquito larvae were used for each assay and four replicates were set up for each concentration and the control (0%). The control beaker contained only 0.5ml of acetone in water. Larval mortality was observed after 24 hours Mosquito susceptibility to botanicals varies across metropolis PLOS ONE | https://doi.org/10.1371/journal.pone.0210440 January 9, 2019 [44]. Larvae that failed to respond to prodding or make it to the surface for respiration were judged to be dead. The assay described above was performed separately using extracts from both botanicals on all the sampled populations.
Fumigant effect of plant extracts on adult mosquito. The fumigant effect of the plant extracts against adult mosquitoes was assessed by using a modified WHO protocol [45] impregnated filter paper technique [46]. The base of a 250ml plastic container (diameter-50mm (upper) and 35mm (lower); height-90mm) was cut open and clogged with cotton wool. The top of the container was covered with muslin cloth fastened with a rubber band. Strips of Whatman's No.1 filter papers (3cm x 3cm) were doused with 0.3 ml of varying concentrations of the plant extracts in acetone (P. guineense [0.7%, 0.9%, 1.2%, 1.4% or 1.6% delivering 0.0084μl/cm 3 , 0.0108μl/cm 3 , 0.0144μl/cm 3 , 0.0168μl/cm 3 or 0.0192μl/cm 3 , respectively] and E. aromatica [6%, 7%, 8%, 9% or 11% delivering 0.072μl/cm 3 , 0.084μl/cm 3 , 0.096μl/cm 3 , 0.108μl/cm 3 or 0.132μl/ cm 3 , respectively]. The solvent was allowed to evaporate from the strip before it was placed below the cotton wool in cut part of the container to avoid contact with the mosquitoes. Ten adult An. gambiae were introduced into each plastic container using a pooter. Each treatment and the control (filter paper treated with 0.3 ml of acetone only) were replicated three times. The assays for the two botanicals were done separately to avoid synergistic effects. Adults were exposed to treatments for 60 minutes before the removal of the treated filter paper and assessment of knockdown. A mosquito was considered knocked down if it lay on its side on the floor of the container and was unable to fly. Mortality was recorded 24 hours post exposure period.

Data analysis
The data were arcsine transformed and subjected to probit analysis [47] to determine the median lethal dose (LD 50 ) and the median knockdown dose (KD 50 ) of both botanical insecticides for each mosquito population. General Linear Modelling (GLM) was used to examine global differences among mosquito populations and treatment means at P < 0.05. Where significant differences occurred, the means were separated using Tukey's post-hoc test. The main effects of location (L) and concentration (C) and their interaction (LxC) on the tolerance of An. gambiae to both mortality and knockdown effect of the botanicals were analyzed.
The relationship between knockdown (KD 50 ) and mortality (LD 50 ), for each population and botanical was investigated using Pearson correlation matrix. Given that different chemical compounds are present in the botanicals, correlation of An. gambiae tolerance between the botanicals for each population was also investigated. All analyses were carried out using Statistical Package for Social Sciences (SPSS) version 20. The tolerance magnitude (T.M.) of the mosquito populations (from each location for each plant extract) was calculated using the following expression: The spatial distribution of T.M. of the mosquito populations was mapped and plotted using ArcMap10.3 software.

Larval LD 50
Irrespective of the plant extract used, generally Alagbaka mosquito populations were the most susceptible, with the lowest LD 50 . The Oke-Aro larvae population had the highest LD 50 when exposed to the P. guineense extract (0.56%) while the Oja-Oba population had the highest LD 50 when exposed to E. aromatica extract (6.09%) ( Table 1). For the homogeneity of response (slope of the log-dose probit relationship), the Oba-Adesida population had the steepest slope for P. guineense (4.12) and E. aromatica (4.27). Shagari population had the shallowest slope for P. guineense (1.50) and FUTA had the shallowest slope for E. aromatica (1.93
Adults. In contrast to the larval results, GLM revealed no significant effect of location (L) on the susceptibility of An. gambiae adults to knockdown and mortality effects of the botanicals. It however showed a significant effect of concentration (C) of the botanicals on the vectors susceptibility to both knockdown (P. guineense: F 5, 118 = 3.19, p = 0.01; E. aromatica: F 5, 118 = 2.39, p = 0.04) and mortality (P. guineense: F 5, 118 = 2.80, p = 0.02; E. aromatica: F 5, 118 = 2.88, p = 0.02). There was no significant interaction between location and concentration on the susceptibility of An. gambiae adults to knockdown and mortality to either botanical.

Relationship between the activities of P. guineense and E. aromatica
The relationships between the activities of the tested botanicals on An. gambiae are shown in Table 4. There were positive correlations between the larvicidal and adulticidal activities of P. guineense (r = 0.14; p = 0.06) and E. aromatica (r = 0.61; p = 0.30) on An. gambiae. However, the correlations were not significant. There were significant positive correlations between adult knockdown and mortality effected by both botanicals (P. guineense: r = 0.81, p = 0.01; E. aromatica: r = 0.82, p = 0.01) ( Table 4).
There was a positive but non-significant correlation between the larvicidal activities of the two botanicals ( Table 4). The adult knock-down activities of the botanicals was also positive but not significant. However, there was a significant positive correlation (r = 0.75; p = 0.01) between the adulticidal activities of the botanicals.

Spatial variation in tolerance magnitude (T.M.)
The spatial distribution of T.M. of An. gambiae larval populations in Akure metropolis to botanical insecticides is shown in Fig 4. In relation to Oke-Aro (the location with the highest T.M.) in the south west of the region, generally low larval susceptibility to P. guineense was observed across the metropolis (Fig 4A). The only exception was seen in the centre of the metropolis where the Oja-Oba population had a T.M. about three quarters that of Oke-Aro (Fig 4A). Similarly, for E. aromatica, a low tolerance magnitude was seen across the metropolis. Mosquito larvae from the centre of the town (Oja-Oba) exhibited the highest tolerance in comparison to the other populations ( Fig 4B). The Oja-Oba An. gambiae adult population had the highest level of tolerance to both the knockdown and mortality effect of the botanicals. This identified Oja-Oba as the tolerance hotspot region (Figs 5 and 6). The FUTA population in the Northwestern part of the metropolis had a KD 50 only one quarter that of the highest KD 50 (Oja-Oba) (Fig 5A) while the Alagbaka and Oda populations in the East and South East, respectively, had KD 50 values only one third that of the Oja-Oba population. Moderate levels of more than half of Oja-Oba tolerance were observed across the metropolis.

Discussion
This study found significant differences in the tolerance of larvae and adult An. gambiae to two botanical preparations with putative insecticidal properties. There are at least two reasons why the susceptibility of larvae and adults is different. Larvae are filter feeders and therefore ingesting the compound, whilst in adults the compounds have to penetrate the insect through the cuticle. Once in the gut the compounds could be actively taken into the body along with nutrients; the compounds are most likely passing passively through adult cuticle. We have no idea how or whether the active ingredients are modified or de-activated differentially between larvae and adults. The second issue is that the compounds are delivered in different ways for larvae and adults. Again, we do not know whether mode of delivery affects the active compounds in any way. Alagbaka populations of An. gambiae were the most susceptible whereas the Oja-Oba populations of An. gambiae adults were the most tolerant to the botanicals. These differences could be due to local environmental conditions. The abundance of susceptible individuals of An. gambiae in an area is sometimes due to the availability of unpolluted breeding sites [49]. It is possible that the presence of susceptible individuals in this area (Alagbaka) is influenced by the availability of unpolluted ground water pools and environmental conditions that encourage rapid development of the mosquitoes. Alagbaka is a Government Reserved Area (G.R.A) in the Akure metropolis. It is a residential area that is characterized by a proper drainage system and waste management. In contrast, Oja-Oba is located centrally in Akure and boasts the biggest market in the metropolis. The wastes or by-product of plant materials are washed into the gutters, some of which are stagnant. Therefore, these water bodies might have been polluted with defensive chemical compounds from these plant sources, thus enhancing the tolerance of An. gambiae to the botanicals investigated. Kim and Muturi [50] examined the relationship between mosquito species reared on leaf litters and induction of cytochrome genes (CYP450). Their result demonstrated that the genes were induced enhancing the metabolism of toxic products that the mosquitoes were exposed to. Other studies have also reported the role of natural xenobiotics in boosting mosquito larvae response to insecticides [51][52]. Considering that the mosquito breeding sites in Oja-Oba contain dissolved plant chemicals or plant particles, the question of the relative impact of these natural xenobiotics on the response of mosquitoes is pertinent [50]. Their presence might affect mosquito metabolism, modifying their tolerance to insecticides. This might increase selection pressure and lead to the development of biocidal resistance. It is commonly found that resistance occurs through the production of detoxifying enzymes that degrade insecticides before they are able to exert their effect [27,[53][54][55]. This phenomenon has been identified in mosquito populations for all major classes of insecticides such as organophosphates and pyrethroids [53][54][55]. Potent botanicals present in an environment due to indiscriminate discarding (as in Oja-Oba) or use for agricultural, economic and other associated purposes, could stimulate the development of resistance. Differences between populations could reflect the local breeding sites, so that mosquitoes that bred in water into which plant material fell (allowing their chemicals to leach into the water) could be better adapted to cope with botanicals [56]. In Southwest Nigeria, as elsewhere, vector control programmes rely exclusively on the application of chemical insecticides, especially pyrethroids, either through the use of spraying or in insecticide treated nets. The control of agricultural and urban insect pests in this region also relies heavily on both synthetic pyrethroids and organophosphates. Nkya et al [33] argued that the presence of natural xenobiotics in mosquito breeding sites and pesticide usage in agriculture influences mosquito response to pyrethroids. GLM revealed a highly significant effect of location on the susceptibility of An. gambiae larvae to the two botanicals used. This is in contrast to the adult stage where GLM revealed no locational effect on both knockdown and mortality. However, a closer look at the spatial variation of T.M. shows locational differences in the response of the adults of the various sampled populations of An. gambiae to the botanicals used. This study revealed that areas with similar habitat characteristics had similar tolerance level, thus suggesting that environmental factors might influence the among population tolerance differences noted here. For instance, the Oba-Adesida and Oja-Oba, areas are characterized by dirtier and more polluted breeding sites. Mosquito populations from these sites exhibited higher levels of tolerance to the botanicals investigated. In contrast, the Alagbaka and FUTA populations of An. gambiae were collected from cleaner water and had higher susceptibility levels. Assaying synthetic pyrethroid and organophosphate on the same An. gambiae populations from the studied areas; it was also observed that some of the implicated areas showed some tolerance similarities to what was obtained with botanicals (Gbaye and Oladipupo, pers comm). The differences in the tolerance magnitude of An. gambiae populations to the botanical insecticides could be linked to the condition of their breeding sites. Brittany et al. [57] reported that the conditions experienced by the larval population of An. gambiae play a key role in adult susceptibility to insecticides. According to Aguirre-Obando et al. [58], Aedes aegypti populations sourced from diverse locations in Brazil showed variations in their level of susceptibility to insecticides used. Similarly, Polson et al. [59], reported that Cambodian populations of A. aegypti larvae manifested variable differences in their level of susceptibility to a synthetic insecticide depending on whether or not they had prior exposure to similar chemicals in their environment. Thus, the knowledge of the effects of environmental conditions on immature mosquito development is important for the inference of results of laboratory experiments on mosquito tolerance.
A significant positive correlation between the adulticidal activities of P. guineense and E. aromatica on An. gambiae populations was observed in this study. This implies that an adult population of An. gambiae susceptible to P. guineense might also be susceptible to E. aromatica. The components of these two botanicals are different and there is not a priori reason why we would expect the insects to respond to them in exactly the same way. Variation in the larval breeding site and also in the genetic background of the population could affect the mosquitoes' response. There exists potential variation in the activities of the used botanicals and/or titres of detoxification enzymes in different mosquito populations which have most likely evolved to deal with plant chemicals [60]. There is the need to increase the number of botanical types to ascertain the universality of this observation. There were also positive correlations between adult knockdown and mortality caused by each botanical. This is in line with the study of Norris et al [61], they also observed correlation between An. gambiae knockdown and mortality caused by several botanical oils {which also include clove bud}. Same cannot be said for Aedes aegypti in their study, hence this observation might be species or genus dependent. It implies that the susceptibility of an An. gambiae population to knockdown by the plant extracts could be a predictor of the mortality caused by the same extract. This contradicts the findings of Owusu et al [62] who argued that knockdown by synthetic insecticide is a poor predictor of 24 h mortality. Although our study is on botanicals with a cocktail of (probable) synergistic chemicals, Owusu et al [62] were actually comparing the 24h mortality in WHO assay with knockdown in the CDC (Center for Disease Control and Prevention) bottle assay. The mix of several active components in botanicals, such as P. guineense and E. aromatica, might have prevented the recovery of several of the An. gambiae populations after knockdown in our study. Feng and Isman [63] argued that a cocktail of active components from the whole of a botanical insecticide is able to deter resistance development better than a single isolated plant compound.
Larvicidal activity of E. aromatica against An. gambiae was observed to have a significant positive correlation with its adulticidal activity unlike P. guineense (positive but relationship not significant). This indicates that any population susceptible to E. aromatica at the larval stage might also be susceptible at the adult stage. Such a relationship is not found with all synthetic insecticides, for example bendicarb (a carbamate) where a negative relationship between An. gambiae larval and adult mortality was found [64].
The centre of the metropolis studied seems to be the hotspot of tolerance to the botanicals investigated. This might be an indicator of probable sites of control failure with any future use of botanicals. There is a need for further work on populations from these locations with regard to enzymatic activity to further corroborate the findings in this study. Although a standardized method of testing botanical efficacy remains to be settled on [65], there is a clear need for wider research to establish the level of tolerance to known botanicals. Likewise, further investigation on the influence of environmental factors on the response of insects to botanicals is paramount to enhance and sustain the development and application of botanical insecticides.

Conclusion
This study revealed that (1) in areas within the metropolis with similar habitat characteristics, An. gambiae had similar tolerance level to botanicals; (2) the susceptibility of an An. gambiae adult population to knockdown by a botanical could be a predictor of the mortality caused by the same plant; (3) a population of An. gambiae susceptible to E. aromatica at the larval stage might also be susceptible at the adult stage; (4) spatial analysis implicated the centre of the metropolis studied to be the hotspot of tolerance to the botanicals investigated, hence, an indicator of probable site of control failure with the future use of the botanicals.
Supporting information S1 File. Data points for An gambiae botanicals. (XLSX)