Malaria vectors in the Democratic Republic of the Congo: the mechanisms that confer insecticide resistance in Anopheles gambiae and Anopheles funestus

Entomological surveys and collections were conducted at the Kibali Gold Mine in the Moto goldfields, north-eastern DRC (N03°08.846′E29°36.548′) in March/April 2011 and July 2012. Adult collections were done by manual aspiration in and around mine houses and structures, as well as in the homes and structures of nearby communities. According to initial field identifications using morphological keys [43], An. gambiae complex and An. funestus group were collected. A sub-set of wild caught females of each group were used for standard WHO insecticide susceptibility bioassays while the remainder of the females (2011: An. gambiae complex = 101, An funestus group = 101; 2012: An. gambiae complex = 88, An. funestus group = 94) were brought back to the insectary at the National Institute for Communicable Diseases (NICD) in Johannesburg for further analysis. For this purpose, females were separated into glass vials, blood fed and provided with moist filter paper for oviposition. Each family produced from a female was given a number for identification purposes.

Mosquitoes were identified to species level by polymerase chain reaction (PCR) for the An. gambiae complex [44], An. coluzzii (previously M molecular form) or An. gambiae s.s. (previously S form) [45], and the An. funestus group [46].

Enzyme-linked immunosorbent assay (ELISA) was used to detect the presence of P. falciparum [47, 48]. The head and thorax of individual mosquitoes were used, and in all assays, positive and negative controls were included. Reactions were assayed spectrophotometrically at 405 nm (Ascent Multiskan RC vl. 5.0, Genesis version 3.03, Labsystems) and all positive samples were assayed a second time.

Where possible, field collected adults (males and females) of each species were exposed in the field to DDT (4%), deltamethrin (0.75%), propoxur (0.1%), fenitrothion (1%), malathion (5%), bendiocarb (0.1%) and dieldrin (4%), according to standard WHO procedures [49]. In each case, mosquitoes were exposed to a particular insecticide for 1 h (2 h for fenitrothion), and sample sizes per exposure varied from 12 to 30 mosquitoes per tube depending on mosquito numbers. Mosquitoes were provided with a 10% sugar solution ad libitum and after 24 h, mortality was recorded. During the 2012 collections, too few An. gambiae were collected for field-based bioassays. For this reason, live specimens were returned to the laboratory and F1 progeny were used for insecticide exposures to DDT, deltamethrin, propoxur, fenitrothion, malathion, bendiocarb and dieldrin.

In the second survey (2012), F1 offspring from individual families generated from both An. funestus and An. gambiae were used for synergist assays prior to deltamethrin exposure. These exposures comprised 1–4 day old males and females. Exposures were carried out as described above, however the number of replicate exposures varied for each family depending on the number of offspring available. These exposures were prepared in conjunction with synergist assays where the monooxygenase inhibitor, piperonyl butoxide (PBO) and the esterase inhibiter, triphenylphosphate (TPP) were used to synergize the F1 offspring of An. gambiae (PBO N = 282; TPP N = 188) and An. funestus (PBO N = 190; TPP N = 182) prior to insecticide exposure. Between 5 and 20 adults were exposed, per replicate, to 4% PBO or 10% TPP for an hour depending on mosquito availability, and then immediately exposed to 0.75% deltamethrin for an hour before being returned to a holding tube. Mortality was recorded after 24 h. Insecticide exposure versus synergist plus insecticide exposure were analysed using the Student’s t-test. An outline of the key assays is provided in Fig. 1.

Both target site mutations and metabolic mechanisms were evaluated for the 2012 field season. Target site molecular markers were evaluated for the following alleles: Rdl (A296G), kdr (L1014S and L1014F), ace-1 ^R (G119S). Known metabolic genes (GSTS1-2, GSTE2, TPX2, CYP6M2, CYP6P1, CYP6AG2, CYP4G16, CYP9L1, CYP6Z1 and SOD1) were evaluated in F1 progeny using quantitative real-time PCR (qPCR).

Polymerase chain reaction (PCR) followed by direct sequencing of the relevant gene was performed in order to test for the presence of kdr east and west mutations in specimens assayed against DDT in the An. gambiae field samples. DNA was extracted from both survivors (N = 47) and mortalities (N = 8), stored on silica gel, using the prepGEM Insect Kit (ZyGEM). Each PCR reaction (containing 2.5ul 10 × PCR Buffer, 1.5 μl MgCl₂ [25 mM], 2.5 μl dNTP’s [2.5 mM each], 3 μl each of AGD1 and AGD2 primers [10 µM], 0.2 μl Takara Taq and 11.3 μl nuclease free water) was subjected to the following cycling conditions: 94 °C/2 min (min), (94 °C/30 s [sec], 50 °C/30 s, 72 °C/30 s) × 40 cycles, and a final extension step at 72 °C for 5 min. The primers, AGD1 (5′ ATA GAT TCC CCG ACC ATG 3′) and AGD2 (5′ AGA CAA GGA TGA TGA ACC 3′), span the region containing the mutations [15]. Results were viewed by electrophoresis and column purified amplicons were sequenced in both directions by Macrogen (Sanger sequencing using the ABI 3730XL DNA Sequencing system). Sequences were aligned using DNASTAR (Lasergene Megalign 2007) software and were screened for the kdr. Positive and negative controls in the form of laboratory strains were included in order to assist with interpretation.

The presence of the A296G mutation was assayed using TaqMan^® technology according to the method of Bass et al. [50]. A mutant specific probe and wild type probe were included in the assay as well as two standard forward and reverse primers. Both survivors (N = 1) and mortalities (N = 32) were assayed and DNA was extracted using the prepGEM Insect Kit (ZyGEM). The resistant laboratory strain COGS (An. gambiae s.s. colonized from the Republic of the Congo in 2009 [18]) was used as a positive (homozygous resistant (RR) and heterozygous (RS) control while the susceptible SUA strain (An. coluzzii from Liberia) was used as a homozygous susceptible control (SS).

PCR of the ace-1 ^R mutation and subsequent digestion was performed in order to detect the G119S mutation in the acetylcholinesterase neurotransmitter [27]. DNA was extracted using the prepGem Insect Kit (ZyGEM) from mosquitoes that both survived (N = 22) and died (N = 25) following exposure to bendiocarb (0.1%). Amplification of the ace-1 ^R gene to produce a 541 base pair product was performed using the forward Ex3Agdir (5′ GAT CGT GGA CAC CGT GTT CG 3′) and reverse Ex3Agrev (5′ AGG ATG GCC CGC TGG AAC AG 3′) primers under the following conditions: 94 °C/3 min, 35 cycles of [94 °C/30 s, 62 °C/30 s, 72 °C/20 s] and a final extension at 72 °C/5 min. Each reaction comprised 2.5 μl 10 × PCR Buffer, 2.5 μl dNTPs [2.5 mM each], 1.5 μl MgCl₂ [25 mM], 4 μl of each primer at a concentration of 2 μM, 0.2 μl Takara Taq and 12.8 μl nuclease-free H₂O. Digestion using Alu I (Promega) restriction enzyme was carried out and visualized with positive and negative controls by agarose gel electrophoresis. Restriction enzyme digestion produces two fragments of 403 and 138 bp in the SS genotype, while three fragments (of 253, 150 and 138 bp) characterize the RR genotype.

RNA was extracted from 1 to 5 day old An. gambiae F1 progeny using TRI^® Reagent Solution (Sigma-Aldrich) [51]. For each of the three biological repeats, 2 mosquitoes from four different families were used (N = 8), and for each biological repeat, four new families were used. A susceptible laboratory strain of An. coluzzii, SUA, was used as a control and three corresponding biological repeats were prepared. The RNA samples were reverse transcribed into cDNA using the QuantiTect^® Reverse Transcription Kit (Qiagen) according to supplier instructions.

Real-time qPCR was used to evaluate transcription of 10 genes commonly known to, or previously implicated in playing a role in insecticide resistance against pyrethroids. Reference gene (RG) selection was performed as specified by the minimum information for publication of quantitative real-time PCR experiments [52] where the expression in 4 potential reference genes, namely ribosomal protein L19 (RPL19), ribosomal protein S7 (RPS7), β-actin and 18S was evaluated and analysed using the MS Excel add-in, NormFinder [53]. For An. gambiae, RPL19 and β-actin showed the most stable transcription in both males and females. The over-transcription of the genes of interest was measured relative to these two RGs. Each qPCR reaction was set up as follows: 12.5 μl IQ™ SYBR super-mix (Bio-Rad), 4 μl primer (concentration optimized for each gene), 1 μl cDNA (100 ng/μl) and nuclease free water to a volume of 25 µl. PCR was carried out using the Bio-Rad CFX96™ Real-Time PCR Detection System with the following cycling conditions: 93 °C/3 min, followed by 35 cycles of [94 °C/20 s, optimized annealing temperature/25 s, 72 °C/30 s] with a single extension at 72 °C/10 min. and a final melt. Standard curves were prepared by 2-fold dilutions of cDNA derived from the wild mosquitoes. For each gene, three biological repeats were assessed, and for each biological repeat, three technical repeats were included for each reaction. Data were analysed using the Pfaffl [54] method. Initially, column-purified PCR product for each gene of interest was sent to Macrogen for Sanger sequencing in both directions in order to confirm (in addition to melt curve analysis) that the correct product was amplified in each case. The genes evaluated in male and female An. gambiae were GSTS1-2, GSTE2, TPX2, CYP6M2, CYP6P1, CYP6AG2, CYP4G16, CYP9L1, CYP6Z1 and SOD1. Primer sequences are provided in Table 1.

As no kdr has been found in An. funestus to date, this was not evaluated. The presence of ace-1 ^R or rdl mutants was not included as An. funestus from this population were susceptible to bendiocarb and dieldrin, respectively.

RNA was extracted from 1 to 5 day old An. funestus F1 progeny [51]. For each of the three biological repeats, two mosquitoes from four different families were used (N = 8), and for each biological repeat, four different families were used. A susceptible laboratory strain, FANG (An. funestus colonized from Angola in 2002), was used as a control and three corresponding biological repeats were prepared. RNA was extracted using the TRI^® Reagent Solution (Sigma-Aldrich) according to supplier instruction, with a DNase treatment included (RNase-Free DNase Set, Qiagen). The quality and quantity of the RNA was assessed using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific Inc.). Four microarrays were prepared using the Agilent 4 × 44 K platform (AMADID 048099)—one for each biological repeat, and a dye swap to account for bias in dye binding. Each array consisted of 33,022 An. funestus specific probes available in Genbank and 1417 Agilent control features. A total of 303 nucleotides belonged to the cytochrome P450 class, 11 to the GST class, 23 to the esterases class and a large number of other insecticide resistant genes.

Labelling was done using the Low Input Quick Amp Labelling Kit (Agilent Technologies) according to manufacturer’s instructions. A starting quantity of 1 μg RNA was incorporated into each labelling reaction. Labelled targets were purified using the RNEasy^® Mini Kit (Qiagen) and successful labelling was evaluated using a NanoDrop Spectrophotometer (Thermo Fisher Scientific Inc.). Hybridization was performed provided dye incorporation was above 0.1 pmol/μl and cRNA yield above 36.2 ng/μl. In addition, specific activity was determined as (concentration of Cy3/Cy5)/(concentration of RNA)*1000 = pmol Cy3 or Cy5 per ug RNA. Specific activity above 8 pmol Cy3 or Cy5 per μg RNA, with a yield above 825 ng was acceptable. Hybridization was performed using the Agilent Gene Expression Hybridization Kit (Agilent Technologies) and was incubated at 65 °C/18 h. After this period, slides were washed and scanned using the GenePix 4000B scanner (Molecular Devices, USA) where the photomultiplier tube (PMT) settings were adjusted to give a pixel ratio of approximately 1. Spot quality and background intensities were examined and corrected using GenePix Pro 6.0 software (Axon Instruments, USA). Saturated features were excluded from analysis (the PMT gain was set at maximum and minimum limits of 63,000).

Gene expression data were analysed using the limma (linear models for microarray data) package version 2.12.0 (Bioconductor) [55] in R version 2.8.0 [56]. Data were normalized (‘within-array’ global loess normalization and ‘between-array’ Aquantile normalization), and linear models were fitted in order to contrast resistant An. funestus expression values with expression values of a susceptible strain. Differentially expressed probes were defined as those with a fold-change greater or equal to 2 and with adjusted p-value ≤ 0.05 [p-values associated with the moderated t-test were corrected for multiple testing based on the false discovery rate (FDR)]. Batch Entrez [57] was used to retrieve the nucleotide sequences linked to accession numbers of all probes on the microarray (25,404 sequences were retrieved in FASTA format). Blast2GO [58] was used to functionally annotate the probe sequences, and for gene ontology (GO) enrichment of over- and under-transcribed gene sets. The p-values for the GO-enrichment analysis were calculated using Fisher’s Exact Test and an adjustment for multiple testing was performed to control the FDR. In order to obtain additional annotation information, basic local alignment search tool (BLAST) [59] searches were carried out against all PEST and Aedes aegypti transcripts (FASTA files were downloaded from Vectorbase [60]). For each An. funestus query sequence, the top PEST and Aedes aegypti nucleotide BLAST hit was recorded if the e-value cutoff was less than 1e-05 (see Additional file 1).

Real-time PCR was used for validation of An. funestus microarray data as described above. Reference genes used for An. funestus were RPL19 and RPS7 as they were found to be the most stable RGs. The genes selected for microarray validation in CYP6M7 and CYP6P9b, and primer details are summarized in Table 2.

Anopheles funestus gene expression arrays were prepared to evaluate changes in gene expression in resistant female An. funestus from the DRC using a susceptible laboratory strain as a reference. According to our criteria for cut-off (FC ≥ 2.0 and adjusted p-value ≤ 0.05), a total of 409 and 280 probes were significantly over- and under-transcribed respectively (Additional file 1). Using Blast2GO, enriched gene ontology terms were assigned to the sets of over- and under-transcribed genes. As expected, genes with a variety of functions produced significantly increased transcript abundance, although proteins with oxidoreductase activity, and cuticular proteins predominated (p-value < 0.05, FDR < 0.1) as determined by enrichment analyses (Table 6). Amongst the most highly over-transcribed genes were ubiquitin c (EZ916542), a component of the ubiquitin proteolytic system [61], at 8.4-fold over-transcribed; troponin C (EZ915656), an important protein in calcium dependent regulation of muscle contraction [62] at 5.7-fold over-transcribed; and the oxido-reductase enzyme sorbitol dehydrogenase, 8.2-fold over-transcribed (Table 7). Other highly transcribed non-detoxification genes included chymotrypsin 1 (7.3-fold over-transcribed), and a range of cuticle proteins. Five known cytochrome P450s were significantly over-transcribed, namely CYP6M7 (7.7-fold), CYP6P9b (3.3-fold), CYP6P9a (2.0-fold) and the duplicated genes, CYP6P4a and CYP6P4b, at 2.0- and 2.1-fold over-transcribed respectively (Table 8). The DDT-metabolizing GSTe2 was 2-fold over-transcribed, while a second GST, un-annotated in An. funestus, but which mapped to GSTs1 in An. gambiae was also significantly over-transcribed (3.6-fold) (Table 8).

Interestingly, a number of immunity related genes were found to be under-transcribed. These included serine proteases, including CLIPC7 and CLIPB1, a spondin, and trypsin and chymotrypsin like protein. In addition to these analyses, the microarray reporter sequences obtained here were blasted against An. gambiae (PEST) and Aedes aegypti genomes in order to supplement the annotations as the An. funestus reference genome is incomplete [63]. These annotations have been added to the Additional file 1.

Validation of microarray data was carried out on CYP6P9b and CYP6M7. The FC values obtained for these were in line with those obtained by microarray analyses. Specifically, CYP6P9b showed 5.8 (± 1.5)-fold higher expression relative to FANG, and CYP6M7, 6.3 (± 4.2)-fold higher expression than that of FANG. These values are based on the use of two RGs in each case.