Researchers at LSTM have identified a completely new mechanism by which mosquitoes that carry malaria are becoming resistant to insecticide.
After studying both Anopheles gambiae and Anopheles coluzzii, two major malaria vectors in West Africa, they found that a particular family of binding proteins situated in the insect’s legs were highly expressed in resistant populations.
First author on a paper published today in the journal Nature, Dr. Victoria Ingham, explains: “We have found a completely new insecticide resistance mechanism that we think is contributing to the lower than expected efficacy of bed nets.
The protein, which is based in the legs, comes into direct contact with the insecticide as the insect lands on the net, making it an excellent potential target for future additives to nets to overcome this potent resistance mechanism.”
Examining the Anopheline mosquitoes, the team demonstrated that the binding protein, SAP2, was found elevated in resistant populations and further elevated following contact with pyrethroids, the insecticide class used on all bed nets.
They found that when levels of this protein were reduced, by partial silencing of the gene, susceptibility to pyrethroids were restored; conversely when the protein was expressed at elevated levels, previously susceptible mosquitoes became resistant to pyrethroids.
The increase in insecticide resistance across mosquito populations has led to the introduction of new insecticide treated bed nets containing the synergist piperonyl butoxide (PBO) as well as pyrethroid insecticides.
The synergist targets one of the most widespread and previously most potent resistance mechanisms caused by the cytochrome P450s. However, mosquitoes are continually evolving new resistance mechanisms and the discovery of this new resistance mechanism provides an excellent opportunity to identify additional synergists that could be used to restore susceptibility
Professor Hilary Ranson is senior author on the paper. She said: “Long-lasting insecticide treated bed nets remain one of the key interventions in malaria control.
It is vital that we understand and mitigate for resistance within mosquito populations in order to ensure that the dramatic reductions in disease rates in previous decades are not reversed. This newly discovered resistance mechanism could provide us with an important target for both the monitoring of insecticide resistance and the development of novel compounds able to block pyrethroid resistance and prevent the spread of malaria.”
Malaria is a life-threatening disease causing more than 500 000 deaths annually in sub-Saharan Africa, mostly in children under five and pregnant women. Prevention of the disease is best achieved by vector control which, nowadays in Africa, largely relies on contact insecticides.
However, the evolution and spread of insecticide resistance seriously threatens the success and sustainability of control interventions, and it is possibly associated with the recent rise of malaria cases in 2018, for the first time since 2000.
Overexpression of detoxification enzymes, which inactivate or sequester insecticides, and mutations in the target site that alter the affinity of insecticide binding have been widely described in malaria vectors, however, may not explain the recent emergence of striking multiple-resistant phenotypes in West Africa [1].
Insecticides must first bypass both cellular and non-cellular leg barriers to reach their targets within the neuronal cells, the voltage-gated sodium channels [2,3]. Mosquito legs play a major physiological role in their life. They have been adapted to implement diverse biological processes, ranging from walking on top of water or downwards on ceilings to sensing mechanical or chemical stimuli [4].
The outermost part of the mosquito leg is the cuticle serving a variety of functions like protection from desiccation, chemical communication and sensory perception of the environment, mechanical support and locomotion. The leg cuticle is also the first barrier protecting the insect from the penetration of external compounds, such as insecticides [5].
The cuticle is sub-divided into two layers: the ‘inner’ procuticle, which forms the bulk of the cuticle, and the outermost layer, called the epicuticle. The procuticle contains chitin filaments complexed with cuticular proteins (CPs) [6,7], but how exactly these proteins interact to each other or with chitin remains elusive. A plethora of CPs belonging to diverse protein families (CPR, CPAPn, CPG, CPF, CPLCG and more) have been identified in many insects so far, including mosquitoes [8].
The majority of CPs belong to the CPR family, with the characteristic Rebers and Riddiford (R&R) Consensus that confers extended chitin-binding properties [7]. Recent studies show that CP members contribute to cuticle maintenance and structure, since silencing of specific genes resulted in thinner and/or malformed cuticles [9,10].
The epicuticle surface is covered by lipids, the majority of which are hydrocarbons (CHCs), free fatty acids and wax esters, while highly tanned proteins associated with lipids are located in inner layers [11–13]. CHCs are the most abundant lipid species in the Anopheles epicuticle, forming a complex mixture of multi-isomeric methyl- and dimethyl-branched components, as well as straight-chain alkanes or monounsaturated alkenes [14].
The final step of CHC production is catalysed by the insect-specific CYP4Gs in specialized cells, the oenocytes, located in abdominal walls [15,16].
Upon biosynthesis, large amounts of CHCs are bound to lipophorins in the haemolymph and are finally transported to different body tissues, such as oocysts, as well as being transferred for deposition on insect surface [17–21]. The exact deposition mechanism, including uptake and travelling through epidermal cells and integument, remains unknown. On top of the cuticle, CHCs form a highly waterproofing layer, which confers desiccation resistance [22–24].
Moreover, different CHC blends are characteristic for different insect species, whereas the species-specific CHC signature may vary quantitatively because of different geographical origins [25]; they can also discriminate age and/or sex within species [26–30], as well as playing dual roles in desiccation resistance and pheromonal communication [31].
A role of cuticle thickening in conferring insecticide resistance, one of the most urgent WHO concerns for malaria control, has previously been demonstrated. A strain of An. gambiae originating from West Africa, resistant to pyrethroids and DTT, had thicker procuticle in the femur leg segment, a phenotype correlated to over-expression of CPLCG3 and CPRs [32].
In Culex pipiens pallens, the femur cuticle was also thicker in resistant versus susceptible populations, mainly because of CPLCG5, since gene-specific silencing resulted in a thinner cuticle and increased susceptibility upon insecticide contact [10,33].
Additionally, a multi-resistant An. gambiae mosquito strain from Côte d’Ivoire (Tiassale) had a thicker femur cuticle (figure 1a), primarily due to enriched deposition of CHCs, which delayed the penetration rate of contact insecticides, and thus presumably produced a more intense and broader insecticide resistance phenotype [14,34]. Noteworthily, CHCs were enriched in Tiassale-resistant An. gambiae, compared to a susceptible population, at the whole-body level, a difference attributed primarily to quantitative, but not qualitative changes [14].
All the above cases report cuticular thickness changes in femurs, although these leg compartments do not come in direct contact with insecticide molecules per se. Insecticide contact would be mediated predominately by the tarsi, and hence these most distal leg segments are expected to play the most important role in insecticide uptake (figure 1a). However, changes in this leg compartment have not been studied to date.
So far, cuticle alterations have been associated with insecticide resistance in An. gambiae and other insects, but the cellular and molecular underpinnings of this phenomenon remain largely unclear. Transcriptomic analysis indicated the association of elevated expression of both cuticle genes and genes associated with the hydrocarbon biosynthetic pathway (such as CYP4Gs, elongases, fatty acid synthetases), with the resistance phenotype in An. gambiae and An. arabiensis [35,36], while the role of CYP4G16 in the biosynthesis of An. gambiae CHCs was functionally confirmed [14]. However, the detailed mechanism responsible for lipid transport and epicuticular deposition remains unknown.
Here, by using transmission electron microscopy (TEM), and lipidomic and proteomic approaches, we have studied structural and functional alterations in the legs of highly resistant An. gambiae, compared to susceptible counterparts.
Material and methods
(a) Mosquito strains
The An. gambiae N’Gousso (NG) strain from Cameroon is susceptible to all classes of insecticides, whereas the An. gambiae VK7 strain from Burkina Faso (VK7) is resistant mainly to pyrethroids and DDT [37]. Mosquitoes were reared under standard insectary conditions at 27°C and 70–80% humidity under a 12 h : 12 h photoperiod. Both strains have been maintained overtime under the same laboratory conditions before analysis.
(b) Transmission electron microscopy
The cuticle thickness of mosquito legs was measured by TEM, as previously described [14]. Before TEM, one wing from every individual was removed, measured and used as proxy for body size [38]. Individuals with similar wing size were selected and further analysed by TEM. Ultra-thin gold sections of the tarsi, the most distal leg segments immediately after the two claws, were taken from groups of 25 female mosquitoes (resistant and susceptible), 3–5 days old, and observed under a high-resolution JEM 2100 transmission electron microscope (JEOL), at an operating voltage of 80 kV.
Raw images from electron microscope were analysed in Image J (v. 1.52e) and the results from at least five independent measurements are presented as means ± standard error (±s.e.). The statistical analysis of cuticle thickness of both resistant and susceptible legs was performed by SPSS software tool v. 22 for Windows (SPSS Inc., Chicago, IL, USA). Differences in the cuticle thickness between the two samples (resistant and susceptible) were regarded significant (ANOVA, p < 0,001).
(c) Extraction of lipids and cuticular hydrocarbon fractionation, identification and quantitation
Cuticular lipids from 5-day-old-females and males both from resistant and susceptible populations (non-blood fed)—female resistant (FR), male resistant (MR), female susceptible (FS) and male susceptible (MS)—were extracted by 1-min immersion in hexane (x3) with gentle agitation; extracts were pooled and evaporated under an N2 stream (10 insects per tube, n = 7 for each group of mosquitoes). CHCs were separated from other components by adsorption chromatography on a minicolumn (2.5 × 0.5 cm i.d.) of activated SUPELCOSIL A (Supelco), eluted with hexane (4 ml) and then concentrated under an N2 atmosphere.
A similar procedure was used for the extraction and fractionation of leg CHCs (60 legs per tube, n = 7 for each group of mosquitoes). All seven biological replicates were collected from three serial mosquito generations. CHC identification by gas chromatography-mass spectrometry (GC-MS) and CHC quantitation by GC-flame ionization detector (FID) were performed as described previously [14,39].
Statistics were analysed using GraphPad Prism software, v. 5.03. Differences in the total CHC values were analysed with both Student’s t-test and one-way ANOVA.
(d) Classification and principal component analysis analysis of cuticular hydrocarbons
Relative abundance of 70 CHC species was measured for 14 female mosquito samples (pools of 10 individuals per sample), and finally 35 CHC species (out of 70) that satisfied the abundance criteria (greater than 0.1%) were selected. Seven samples originated from the VK7 (resistant) mosquito population and the remaining seven from the NG (susceptible) mosquito strain. CHCs were isolated from two different body parts from each mosquito pool: the legs and the rest of the body (excluding legs). In all, our dataset consisted of 35 CHC measurements assigned to (a) body part (legs or bodies) and (b) susceptibility (susceptible or resistant).
Abundances were normalized through quantile normalization and were initially subjected to an exploratory PCA analysis using different combinations of target variables (including susceptibility and body parts), which suggested the existence of particular CHCs that could be used as discriminatory features for body part of origin and susceptibility in female samples. Subsequently, a random forest classification strategy was employed (electronic supplementary material) and the abundances of resulting CHCs were subjected to the final PCA analysis.
(e) Comparative proteomics analysis by nLC-MS/MS
Leg protein extracts from both mosquito strains were prepared using the solubilization buffer (SB: 8 M Urea, 0.2 M NaCl, 0.1% SDS, 50 mM Tris–HCL, pH 8.0). Briefly, legs from 25, 3–5-day-old-female and non-blood fed mosquitoes, were dissected and mechanically homogenized in SB by hand pestles. Upon centrifugation (2500 g for 15 min at 10°C), the pellet was discarded and the protein samples were separated on 10% polyacrylamide gels, following staining with 0.1% Coomassie Brilliant blue R-250 (in 40% methanol 10% acetic acid) and destaining with 10% ethanol, 7.5% acetic acid [40]. Each lane was excised and cut in 15 bands, which were subsequently destained, reduced, alkylated and in-gel tryptic digested as previously described [41].
Tryptic peptides were finally dried in a speed-vacuum centrifuge and dissolved in 5% FA in ultra-pure water solution. All samples were desalted by being loaded on conditioned home-made pre-columns packed with C18 extraction disks (Empore) and eluted stepwise with 80% MeOH and 5% FA. All elution fractions were collected, speed-vacuum centrifuged and diluted in 5% FA for further nLC-MS/MS analysis.
Protein identification and relative quantitation by nLC-MS/MS was done on a LTQ-Orbitrap XL coupled to an Easy nLC (Thermo Scientific). The sample preparation and the LC separation were performed as previously described with minor modifications [42].
Briefly, the dried peptides were dissolved in 20 µl 0.5% formic acid, aqueous solution and the tryptic peptide mixtures were separated on a reversed-phase column (ReprosilPur C18 AQ, particle size = 3 µm, pore size = 120 Å, Dr Maisch), fused silica emitters 100 mm long with a 75 µm internal diameter (Thermo Scientific), packed in-house, using a pressurized (35–40 bars of helium) packing bomb (Loader kit SP035, Proxeon). The nLC flow rate was 200 nl min−1. Tryptic peptides were separated and eluted in a linear water-acetonitrile gradient and injected into the mass spectrometer [42,43].
MS survey scans were acquired in the Orbitrap from 200 to 2000 m/z at a resolution of 60 000 and for the MS/MS, precursor isolation at 1.6 m/z was performed by the quadrupole (Q). Fragmentation of 20 most intense ions by collision-induced dissociation (CID) with normalized collision energy of 35% and rapid scan MS analysis were carried out in the ion trap. The dynamic exclusion duration was set to 15 s with 10 ppm tolerance around the selected precursor and its isotopes. The AGC target values were set to 4.0 × 105 and 1.0 × 104 and maximum injection times were 50 ms and 35 ms for MS and MSn scans, respectively.
MS raw data were analysed by Proteome Discoverer 1.4.0 (Thermo Scientific) using Mascot 2.3.01 (Matrix Science) search algorithm. Spectra were run against the latest version of An. gambiae theoretical proteome containing 13 515 entries (UniProt 2018) and a list of common contaminants [44,45]. Search parameters employed are described in detail elsewhere [46]. Final peptide and protein lists were compiled in Scaffold (v. 4.4.1.1, Proteome Software; Portland, OR, USA) employing criteria previously described [46]. Manual annotation of identified proteins was also performed.
To determine the differentially expressed peptides, we compared the abundances of the peptides between resistant and susceptible control mosquito legs, using a minimum of 3.0-fold changes (normalized to median). GO terms and pathways enrichment analysis of the differentially expressed proteins, based on fold change criteria, was performed using Gene Ontology Consortium [47]. In addition, a beta-binomial test, using the ibb library in R, was employed to identify significant changes at single-protein level between the control and resistant samples among the three biological replicates. p-values < 0.1 were considered to be significant.
(f) Western blot analysis
Legs from twenty-five, 3–5-day-old female mosquitoes were dissected and mechanically homogenized in solubilization buffer (SB: 8 M urea, 0.2 M NaCl, 0.1% SDS, 50 mM Tris–HCL, pH 8.0). Upon centrifugation (2500 g for 15 min at 10°C), pellet was discarded and the supernatant was analysed by SDS-PAGE and western blot. Antibody against b-Tubulin (E-10, 1 : 500) was from Santa Cruz Biotechnology (SC-365791), and antibody against CPCFC1 (1 : 500) was the anti-AgamCPCFC1 produced in Vannini et al. 2017 [7] and were kindly provided by J. Willis.
(g) Quantitation of chitin
Chitin determination in mosquito legs was made according to Lehmann & White [48] and subsequent modifications [49–51]. Briefly, legs and remaining bodies from both mosquito strains (susceptible and resistant) in three replicates of eight female mosquitoes, each 3–5 days old (48 legs each replicate), were mechanically homogenized and further processed for deacetylation of chitin to chitosan and determination of chitosan (i.e. glucosamine polymer). Finally, samples were transferred to a 96-well microplate and absorbance was determined at 650 nm in a plate reader (Molecular Devices, Spectra Max). Before chitin quantification, mosquitoes’ leg and body weights were measured and used for normalization.
Discussion
This study describes substantial differences in the properties of the insect cuticle in two strains of An. gambiae that differ in their insecticide resistance profile. Although these strains are from geographically distinct locations and have been maintained in colony for multiple years, the compelling evidence for quantitative differences in the cuticle both in the total amounts as well as in the relative amounts of certain components of the resistant strain used in this study is supportive of previous work demonstrating the importance of reduced penetration through the cuticle as a key resistance mechanism in An. gambiae.
Importantly, this study highlights the critical role of cuticle composition in legs, as the body part that contacts insecticide used in indoor applications for malaria control, and provides novel insights into the putative mechanisms of cuticular resistance.
We have demonstrated that changes in CHC content, as well as structural procuticle components, could be responsible for building a thicker leg cloak in the tarsus of VK7, a highly resistant An. gambiae population from W. Africa. The legs from resistant mosquitoes, in particular, had a substantially higher CHC content, compared to susceptible mosquitoes. A difference in the content of CHCs had been previously reported at the whole body level of another resistant population (Tiassale) from Côte D’ Ivoire [14], but the current study shows that this difference is indeed primarily attributed to changes in the legs, the most physiologically relevant body part.
Moreover, the accumulation of CHC mass in legs of resistant mosquitoes possibly implicates transport mechanisms towards the legs and/or towards the leg’s epicuticle, since CHC biosynthesis probably occurs in the abdominal oenocytes [16]. Thus, local concentration of specific CHC binding molecules like lipophorins and pheromone-binding proteins (PBPs) in the leg’s haemolymph and the expression of their specialized receptors is increased and/or transporter molecules such as ABC transporters expressed in the leg’s epidermis are more abundant or active.
Additionally, we found six methyl-branched CHC peaks as the most differentiated CHCs, in terms of relative amounts, compared to the CHC reservoir of the remaining body, also implicating a selective transport of CHCs to the legs. Selective transport pathways for the export of various CHC components to the epicuticle and to different tissues, have been suggested in many studies [17,52], including specific transport of volatile sex-pheromone in moths and social ants [53,54].
Along with the overall quantitative differences, resistant legs showed a characteristic CHC signature based on their relative concentrations, when compared to the susceptible control legs. Six CHC species were particularly over-represented in the resistant legs compared to susceptible ones. Moreover, CHCs as major components of the insect epicuticle exert a broad spectrum of functions affecting the whole insect’s physiology. Generally, very long-chain alkanes contribute to the waterproofing properties of the cuticle lipid layer and, ultimately, to desiccation tolerance [55].
Very long-chain alkenes and methyl-branched alkanes might increase the chemical information content of the cuticle, since their role in chemical communication in various insect species is well documented (for example 7, 11-heptacosadiene is the main excitatory pheromone in Drosophila melanogaster females). CHCs have been shown to play dual roles in Drosophila, both in pheromonal communication and in tolerance to desiccation [56,57]. Thus, the mating capacity/fitness of resistant mosquitoes might be affected by the over- or down-representation of specific leg CHCs, although this possibility needs to be further investigated.
Subsequently, the leg proteome was analysed by gel-based proteomics. It consisted of 1120 proteins, with energy production and lipid metabolism proteins enriched compared to the Anopheles whole body proteome. Comparative proteomic analysis between legs from insecticide resistant and susceptible An. gambiae mosquitoes showed that the most upregulated proteins were the structural CPs, in agreement with a leg-thickening phenotype. The over-expression of CPs in the cuticle of resistant insects has been also reported in other insects/mosquitoes at the whole organism level and it has been associated with reduced insecticide penetration and thicker cuticles [58,59].
In addition to the large number of substantially elevated CPs, resistant mosquitoes also had substantially elevated chitin levels in their legs (and not in their remaining bodies) which apparently bound to the elevated CPs and reinforced procuticle chitin-rich layers in the legs. Furthermore, upregulation of chs1 transcripts could be correlated with increased production of chitin in legs of resistant compared to the susceptible female mosquitoes. Noteworthily, two CPs, CPR8 and CPR120, were found to be downregulated in the resistant legs, indicating that not only thickening but also qualitative differences could affect insecticide penetration by a yet unidentified mechanism.
The presence and/or differential expression of protein families which have been previously associated with insecticide resistance were also identified (electronic supplementary material, table S3), such as ABC transporters, detoxification enzymes and odorant binding proteins (OBPs). ABC transporters of subfamily G have been implicated in lipid transport to the cuticle epidermis [60]; however, we have identified AGAP003680-PA expressed in leg proteome, an ABC transporter that belongs to subfamily H.
This transporter has been found differentially expressed in a number of insecticide-resistant populations of An. gambiae across Africa [61]. The role of ABCHs in insects remains largely elusive; however, silencing of one of them in Tribolium castaneum (Tc) adults (TcABC-9C) resulted in a reduction of cuticle lipids, suggesting a possible role in lipid transport towards the cuticle [62]. Moreover, the Snu and ABCH-9C transporters from Drosophila melanogaster and Locusta migratoria (Lm) were shown also to participate in the construction of lipid-based barrier of cuticle [63].
Phylogenetic analysis classified LmABCH-9C, TcABC-9C and AGAP003680 in the same group (group C) of ABCH subfamily [64].
A number of detoxification enzymes, such as P450s, hydrolases, carboxyl esterases, glutathione-s-transferases and antioxidant enzymes, were also expressed in mosquito legs. Among them, superoxide dismutase (CuSOD3) and glutathione-s-transferases (GSTe3 and GSTmic1), found significantly upregulated in the resistant legs, are well-known antioxidant enzymes that protect living cells from oxidative damage, including pyrethroid-induced lipid peroxidation [65–67].
Interestingly, seven OBPs, which can bind small organic compounds and possibly act as scavengers for insecticides [68–70], were differentially regulated in the legs of resistant An. gambiae mosquitoes. We hypothesize that the upregulation of two OBPs in the proteomic dataset (i.e. OBP10: 6.33-fold and OBP57: 3.67-fold) might be associated with the increased binding and transport of the extra CHCs found in the resistant legs [71], and/or sequestration of insecticides, although further studies are required to support this hypothesis.
Conversely, the downregulation of five OBPs (OBP13, OBP5470 and three putative D7-related proteins) could possibly indicate an altered chemoreception mechanism in the legs of resistant An. gambiae, for example, via a reduction of hydrophobic channels facilitating the transport of ligands to their receptors, as previously suggested [72,73].
In conclusion, An. gambiae female mosquitoes seem to build a protective leg cloak to fight against insecticides. Reconstruction of thicker and/or altered leg cuticle, via extra deposition of chitin, proteins and CHCs, act in concert to fight, in first line, against insecticides. These changes might be associated with other major physiological functions in An. gambiae life history, such as pheromonal communication and resistance to desiccation, thus affecting fitness and positive selection of insecticide resistance in
More information: A sensory appendage protein protects malaria vectors from pyrethroids, Nature (2019). DOI: 10.1038/s41586-019-1864-1 , https://nature.com/articles/s41586-019-1864-1
