Background
Malaria prevention largely relies on the use of measures, such as long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) [
1]. Of the four insecticides classes used in public health, pyrethroids are by far the most widely used [
1]. During the past decades, overreliance on pyrethroids in public health and agriculture, led to rapid expansion of pyrethroid resistance in malaria vectors populations which now threatens the continued effectiveness of current control efforts [
2]. Resistance to pyrethroids is mainly due to mutations in the knock down genes (
kdr) and metabolic detoxification mechanisms and is largely prevalent in all major malaria vectors [
3‐
5]. Because of rapid spread rate of this resistance across sub-Saharan Africa, effective measures are needed to mitigate its impact. The World Health Organization (WHO) recommends application of insecticides having different mode of action or temporal replacement by different insecticide classes in case of resistance [
6]. Carbamates and organophosphates due to their different mode of action, are actually considered as suitable alternative insecticides to pyrethroids for vector control such as IRS [
7‐
10]. Field experiments conducted across West Africa showed the effectiveness of carbamates and organophosphates against pyrethroid resistant malaria vector populations [
11,
12]. There is, an increasing number of countries which have started introducing the use of carbamate in their national vector control strategy [
7,
9,
10]. However, increasing reports of carbamates resistance in the main malaria vectors across sub-Saharan Africa [
13‐
17], could jeopardize current efforts to implement appropriate resistance management strategies against malaria vectors. Despite current expansion of bendiocarb resistance little is known on mechanisms promoting this resistance in Central African mosquito populations.
In Cameroon, despite efforts made over the past years to control malaria, the disease is still considered, a major threat [
18,
19]. Major vectors in the country display high level of pyrethroid resistance [
20‐
22]. Studies undertaken in the cities of Douala and Yaoundé, reported particularly high prevalence of pyrethroid and DDT resistance in both
Anopheles gambiae and
Anopheles coluzzii [
23‐
25]. The use of insecticide tools for vector control in households, the selective pressure of pollutants in breeding habitats and uncontrolled use of pesticides in small scale urban vegetable farming are all considered to have caused, the fast evolution of insecticide resistance which is now also affecting bendiocarb [
23,
25,
26]. However, the molecular basis of carbamate resistance remained uncharacterized in
An. gambiae populations in Cameroon. Such information is crucial to guide the implementation of appropriate resistance management strategies to prolong the effectiveness of carbamates in Cameroon. The main resistance mechanisms to carbamates involved metabolic resistance and target-site resistance. Metabolic resistance to carbamates is often conferred by the up-regulation of detoxification genes such as cytochrome P450 s [
27,
28] or carboxylesterases [
29‐
31].
Target-site resistance to carbamates and organophosphates is conferred by a single point mutation causing acetylcholinesterase inhibition [
32,
33]. The mutation encoded by the
Ace-
1
R
gene induces a substitution from glycine to serine at position 119 (G119S). The G119S mutation has also been recorded in several species including
Culex quinquefasciatus,
Anopheles albimanus and
An. gambiae [
33‐
38]. Recent findings reported the duplication of this mutation in some
An. gambiae individuals [
39,
40]. In Cameroon, no G119S mutation has up to now been reported in
An. gambiae population and the underlying molecular basis of the carbamate resistance in this major malaria vector remain to be established. The present study seeks to characterize mechanisms involved in the ongoing mosquito resistance to carbamate in the city of Yaoundé. The study also traces the dynamics of
An. gambiae susceptibility to bendiocarb between 2010 and 2013.
Methods
Study site
Mosquito collections were conducted in districts of the city of Yaoundé (3°51′N 11°30′E). Yaoundé the capital city of Cameroon, is situated within the Congo-Guinean phytogeographic domain and display an equatorial climate consisting of four seasons: two rainy seasons (March–June and September–November; annual rainfall 1700 mm) and two dry seasons (December–February and July–August).
Mosquito collection
Mosquito larvae were collected at all stages in water collections across the city of Yaoundé and reared separately according to their breeding habitats characteristics classified as cultivated, polluted or non polluted sites. Water collections with organic wastes were considered as polluted, non-polluted breeding sites were water collections without any sign of organic pollution, cultivated breeding sites were water collections associated with farming practices. In the laboratory, larvae were transferred into distilled water and reared separately at room temperature. During this period, they were fed using fish food until the pupa stage. Pupa were collected in cups and placed inside cages covered with netting for emergence.
Insecticide bioassays
Bioassays were conducted from October 2010 to December 2013 using 2–4 days old females emerging from larvae collected on the field. Morphological identification keys [
41] were used to differentiate members of the
An. gambiae complex to other mosquito species at both the larval and the adult stages. Unfed
An. gambiae s.l. females aged 2–4 days were exposed to 0.1 % bendiocarb, 4 % DDT (dichloro-diphenyl-trichloroethane), 0.75 % permethrin, 0.05 % deltamethrin and 4 % malathion in susceptibility test kits from the WHO, following standardized procedures [
42]. For bioassays using piperonyl butoxide (PBO) as synergist, unfed
An. gambiae females were pre-exposed to 4 % PBO papers for 1 h before being immediately exposed for another 1 h to bendiocarb. Mortality was scored after 24 h but for mosquitoes surviving exposition to bendiocarb, they were maintained in observation for a total period of 48 h before storage in RNAlater. For each bioassay, exposition of mosquitoes to untreated papers was also undertaken as controls. Abbot formula [
43] was used to adjust mortality rate in tested samples if the control group mortality rate was 5–20 %. WHO recommendations [
42] were applied for classifying mosquitoes as resistant or susceptible.
Odd ratio calculations were undertaken to assess any association between phenotypes and genotypes [(resistants genotype A*susceptibles genotype B)/(susceptibles genotype A*resistants genotype B)]. Odd ratio estimates, mortality rates, the 95 % confidence intervals and p values were calculated with the software MedCalc V11.5.0.0.
Molecular identification of species and genotyping of Ace-1R G119S mutation
Genomic DNA utilized for the identification of
An. gambiae s.l. species and the screening of the Ace-1
R G119S mutation, was extracted from a leg or wing of adult mosquitoes by the Livak technique [
44]. A polymerase chain reaction (PCR) was used for
An. gambiae species identification [
45]. The presence of the G119 mutation was screened using TaqMan assays as previously described [
46]. TaqMan reactions were undertaken using the Agilent MX3005P machine. Each reaction was conducted in a 10 μl final volume with 1xSensiMix (Bioline), 800 nM of each primer and 200 nM of each probe.
Microarray experiments
Microarray experiments were conducted using only An. gambiae samples originating from cultivated sites where bendiocarb resistance was most prevalent. Differentially transcribed genes were compared between resistant, control (unexposed) and susceptible (Kisumu) samples.
Pools of ten mosquitoes were used for total RNA extraction with the PicoPure RNA isolation Kit (Arcturus, Applied Biosystems, Mountain View CA USA). Each sample was constituted of three biological replicates. Total RNA extracted from mosquitoes was treated using DNase (RNase free DNase set, Qiagen Hilden Germany). A nanodrop spectrophotometer (Nanodrop Technologies UK) and a Bioanalyser (Agilent Technologies UK) were used to assess RNA concentration and quality. After amplification undertaken using 100 ng of total RNA, samples were labelled using Cy-3 or Cy-5 dye with the “Two colors low input Quick Amp labeling kit” (Agilent technologies, Santa Clara, CA, USA). This was immediately followed by samples purification undertaken using Qiagen purification kit. A spectrophotometer (NanoDrop Technologies) and Bioanalyzer (Agilent Technologies) was used to check for cRNA labelling and yield. Labelled cRNAs were hybridized to the ‘
An. gambiae’ array Agilent 8x15 k chip (AGAM_15 K) (A-MEXP-2196) [
46]. After 17 h hybridization at 65 °C and 10 rpm rotation, slides were washed according to the manufacturer instructions (Agilent Technologies). Microarray slides were then scanned with the Agilent G2565 Microarray Scanner System via the Agilent Feature Extraction Software (Agilent Technologies).
Five hybridizations per comparison including three independent biological replicates and two dye swaps were performed. Resistant samples were competitively hybridized against unexposed samples and the Kisumu laboratory strain.
Microarray data analysis
Genespring GX 11.1 software (Agilent Technologies) was used for microarray data analysis. Comparison of genes expression profiles between groups was undertaken after computing the mean transcription expression ratios to a one sample Student’s
t test against zero. Benjamin and Hochberg calculation [
47] was applied for multiple testing corrections. Transcripts significantly and differentially transcribed were those displaying both
t test p values <0.05 and a fold change ≥twofold compared to the control or susceptible group. Gene ontology (GO) enrichment was performed using David functional 6.7 [
48,
49] to determine GO significantly enriched using as background for comparison the totality of genes differentially transcribed for each group.
Microarray validation by qRT-PCR (real-time quantitative reverse transcription polymerase chain reaction)
Quantitative RT-PCR analysis as described in Tene et al. [
24] was used to confirm the overexpression of detoxification genes detected by microarray. Biological replicates consisting of two micrograms of total RNA per replicate were reverse transcribed into cDNA in a reaction mix containing superscript III (Invitrogen, Carlsbad, CA, USA) and oligo-dT20 primer as recommended by the manufacturer. A MX3005 Agilent system (Agilent) was used to perform quantitative PCR reactions. Each reaction was conducted in a final volume of 25 µl containing iQ SYBR Green supermix (Biorad), primers at the concentration of 0.3 µM each and 5 µl of 1:50 diluted cDNA. The specificity of PCR products generated was verified using melt curves analysis. Standard curves for each gene were generated using serial dilutions of cDNA. Selected transcripts fold changes were normalized to EFGM_ANOGA (AGAP009737_RA) and 40S ribosomal protein S7 (AGAP010592_RA). Fold changes differences of selected genes between test samples and susceptible (Kisumu), were estimated according to the 2
−ΔΔCT method considering PCR efficiency [
50,
51].
Discussion
Despite fast evolution of insecticide resistance in vector populations across Cameroon, molecular mechanisms conferring resistance are still poorly studied. The present study was conducted to characterize molecular mechanisms promoting bendiocarb resistance in An. gambiae populations in the city of Yaoundé. Both An. gambiae and An. coluzzii were found resistant to bendiocarb. Mosquitoes originating from cultivated sites were found to be more resistant to bendiocarb than those collected from polluted or unpolluted sites and could be related to their frequent exposition to xenobiotics including insecticides.
No mosquito was found carrying the G119S mutation conferring target site resistance to carbamate and organophosphate. The increase mortality after the use of piperonyl butoxide (PBO) as synergist suggested the likely implication of cytochrome P450 s in bendiocarb resistance. Our data was similar to previous investigations conducted across West Africa supporting the implication of metabolic mechanisms in carbamate resistance [
27]. Although G119S mutation is recognized as the primary resistance mechanism against carbamates and organophosphates it remains less expanded across Central Africa [
54]. Its distribution might be constrained by its high fitness cost [
39]. However, possession of both G119S mutation and metabolic resistance could lead to extremely resistant phenotypes [
27,
40].
Microarray analysis identified several cytochrome P450 genes with the most important being
cyp6z3, cyp6z1, cyp12f2, cyp6p4 and
cyp6ag1, which were overexpressed when resistant or unexposed samples were compared to the Kisumu susceptible strain (R
b-S and C-S). However, in addition to their potential implication in insecticide resistance, the high fold change difference detected for some of the genes could likely results from the different genetic background between Kisumu strain originating from Kenya and local
An. gambiae populations from Cameroon. Similar observations have been reported from previous studies [
24]. The over-expression of the two P450 genes
cyp12f1 and
cyp4c36 in the comparison between bendiocarb resistant and control non exposed mosquitoes (R
b-C) was low and not observed in the R
b-S comparison suggesting that these genes may not be the main bendiocarb resistance genes. Although further functional characterization studies will help to establish the exact role of these candidate genes.
Cyp12f1 gene was already reported overexpressed in mosquitoes resistant to DDT [
53] while no role for
cyp4c36 in insecticide resistance have so far been reported. Nevertheless cytochrome P450 are known to metabolise a large number of xenobiotics including pyrethroids and carbamates [
55,
56]. For the set of genes detected only overexpressed in comparison between control and susceptible (C-S), despite a probable absence of role in bendiocarb resistance, it is likely that these detoxification genes (
cyp6m2, gstms2, tpx2, gsts1-
1) as well as many others detected over-expressed, might be implicated in the metabolism of an important number of compounds since mosquito populations screened during the study were also recorded resistant to DDT and pyrethroids.
Among potential candidate genes conferring bendiocarb resistance,
cyp12f2 was reported over-expressed in response to bacterial challenge or during malaria parasite invasion in mosquitoes [
57] and in permethrin-resistant
An. arabiensis in South Africa [
58].
cyp6ag1, cyp6z3 and
cyp6p4 were reported over-expressed in DDT and pyrethroid resistant
An. gambiae and/or
An. arabiensis populations [
53,
58‐
60]. Ortholog of cyp6p4 and cyp6z3 have been connected to pyrethroid resistance in the malaria vector
Anopheles funestus [
3,
61]. Whereas, cyp6z1 in addition to its confirm involvement in DDT and pyrethroid resistance in
An. gambiae [
62,
63], was recently reported as the main gene conferring metabolic resistance to bendiocarb to
An. funestus the other major African malaria vector [
28].
Previous investigations from Yaoundé identified several candidates genes including
cyp6m2, cyp6p3, cyp6z3, gstd1-
6, involved in DDT or pyrethroid resistance [
24].
Cyp6m2 and
cyp6p3 also emerged as main candidate genes conferring bendiocarb resistance in a study conducted in Côte d’Ivoire [
27]. However, none of these two genes emerged as potential candidate for bendiocarb resistance. The fact that during the present study only
An. gambiae individuals were screened for microarray analysis while in Côte d’Ivoire mosquito population screened consisted exclusively of
An. coluzzii might somewhere explain the difference recorded. Different detoxification gene expression pattern have been recorded for
An. gambiae,
An. coluzzii or
An. arabiensis [
53,
59,
64]. Several Glutathione S transferase genes including
gstms3, gstms1, gstd1-
3 and
gstd7 were also detected overexpressed in R
b-S and/or C-S comparisons. GSTs are known to metabolize several xenobiotics including pyrethroids, organochlorines and organophosphates and to catalyse the secondary metabolism process of a large number of compounds oxidized by cytochrome P450 [
30,
65,
66]. In pyrethroid resistant strains, the overexpression of GSTs attenuates lipid peroxidation induced by pyrethroid and reduce mortality [
67].
In the city of Yaoundé, mosquito tolerance to DDT and pyrethroids and the prevalence of the
kdr allele, have been increasing with time [
25,
68]. It remains to be established whether increase resistance to DDT and pyrethroids could also have promoted cross-resistance to carbamates. Yet the increase prevalence of bendiocarb resistance poses serious challenges for malaria control since carbamates are considered as a main alternative to pyrethroids.
Authors’ contributions
Conceived and designed the study protocol: CAN, CSW, Participated in field and laboratory analyses: EK, RP, BTF, PAA, CAN, CSW. Critically revised the manuscript: CSW, CC, BTF, RP, PAA Interpreted, analysed data and wrote the paper: CAN with contribution of other authors. All the authors read and approved the final manuscript.