Background
Malaria remains a major public health burden, affecting approximately 240 million individuals annually and causing more than 800,000 deaths, mainly in populations living in tropical and sub-tropical countries in sub-Saharan Africa[
1]. To date, the absence of a licensed malaria vaccine[
2‐
4] and the spread of parasite resistance against malaria treatment[
5] necessitate strengthening the control of malaria exposure by avoiding host/vector contact. Thus, several strategies could be used to protect individuals from mosquito bites, either by using personal anti-vectorial devices, like impregnated bed nets, repellents, and long-sleeved clothes[
6], or by controlling vector populations at both the adult and larval stages[
7,
8]. The effectiveness of these anti-vectorial devices is generally evaluated with parasitological and entomological methods[
9‐
11]. Although these methods have demonstrated their capacity to estimate human exposure to malaria vectors and
Anopheles densities, these tools lack important logistics and present limited efficiency in the context of low-level exposure to
Anopheles bites. In addition, they are not designed for the assessment of the heterogeneity of mosquito exposure at the individual level[
9]. Therefore, the development of new indicators and methods to evaluate the effectiveness of anti-vectorial strategies at the individual level is necessary.
Mosquito salivary proteins injected into the host during blood feeding play a dual role by counteracting homeostasis and modulating the vertebrate immune response[
12]. In addition to their role in the blood meal, some salivary proteins presenting immunogenic properties could elicit an antibody response by their host. This immune response, initially described in allergic individuals[
13], has been proposed as a marker of exposure to mosquito bites[
14,
15]. Thus far, several studies have demonstrated that the level of IgG immune responses against salivary antigens is associated with the level of individual exposure to mosquito bites, which may vary according to seasonal mosquito density[
15,
16], transient exposure following travel in malaria-endemic areas[
17] or following the introduction of anti-vectorial measures, such as the use of insecticide-treated nets[
18]. However, the existence of homologous salivary protein sequences that are shared among different species from
Culicidae requires the identification of specific antigenic proteins or peptides prior to developing any anti-saliva based immunological tools to assess individual exposure to different mosquito vectors[
9,
19].
Among mosquito salivary proteins, the
Anopheles gambiae salivary gland protein 6 (gSG6) was proposed as a potential candidate for the examination of specific malaria vector exposure markers[
20]. This small protein, expressed specifically in the salivary glands of adult female mosquitoes, was selected based on its restrictive presence in species belonging to the subgenus
Cellia, including major Afrotropical malaria vectors (e g,
An. gambiae species complex,
Anopheles funestus)[
21], and its immune recognition by individuals exposed to
Anopheles[
17]. To limit production costs, Poinsignon and colleagues designed a gSG6-based peptide sequence (gSG6-P1) according to its predicted immunogenic properties[
20]. The gSG6-P1 peptide was repeatedly reported to be a relevant
An. gambiae-specific marker of exposure[
20‐
23]. Moreover, the high level of amino-acid conservation between gSG6-P1 peptide sequences from
An.gambiae and
An. funestus indicate the potential of this peptide to be an indicator of exposure to both of these main vectors of
Plasmodium falciparum in Africa[
24]. Similar observations were obtained using recombinant forms of whole SG6 orthologs from
An. gambiae and
An. funestus[
25,
26], which could be attributed to the high level of identity among them (i.e., 80%). More recently, it was reported that the level of IgG against gSG6 was positively linked to the risk of malaria pathogen transmission[
27]. Thus, SG6 proteins are currently the best and uniquely relevant indicators of exposure to Afrotropical malaria vectors.
However, the identification of salivary antigenic candidates capable of discriminating individual exposure at the species level could improve the development of this type of immunological test by determining the anopheline fauna biting population. Effectively, together with
An. funestus, mosquitoes from the
An. gambiae s.l. (
An. gambiae s.s. and
Anopheles arabiensis) are the most common vectors of human malaria in sub-Saharan Africa[
28]. These highly anthropophilic
Anopheles species could geographically co-inhabit most sub-Saharan countries[
29]. Malaria parasites can thus be transmitted by multiple and often sympatric vectors[
30‐
32]. However, anopheline fauna could be spatially and temporally influenced by several factors, such as environmental conditions that could seasonally modify the anopheline species proportions and densities. During the dry season, the
An. gambiae s.l. density is decreased for the benefit of
An. funestus. The maintenance of malaria transmission at several sites could be attributed to the presence of
An. funestus[
33]. Thus, in addition to the use of salivary exposure markers to estimate the individual level of exposure to Afrotropical malaria vectors, the characterization of new species-specific anopheline salivary antigenic candidates could be useful for determining the predominant mosquito populations of a study area. Such information could be useful for the adaptation of vectorial control measures against specific mosquito populations or for the estimation of the risk of malaria transmission or persistence.
Thus, the aim of this study was to assess anopheline salivary proteins that could be used as species-specific exposure biomarkers to distinguish An. funestus exposure from An. gambiae s.l. exposure. First, SG6 salivary proteins from An. gambiae s.s. (gSG6) and An. funestus (fSG6) were produced in recombinant forms and evaluated on sera from individuals that were either un-exposed to Anopheles or exposed predominantly to An. funestus or An. gambiae s.l., to confirm that these salivary proteins could be used to detect a predominant exposure to either of these two mosquito species. In addition, recombinant forms of 5′nucleotidase salivary proteins from An. gambiae s.s. (g-5′nuc) and An. funestus (f-5′nuc) were tested on the same sera to assess their potential as species-specific indicators of exposure. The specificity of the IgG response against these selected salivary proteins at the genus or species levels was analyzed by ELISA using sera from individuals living in three Senegalese villages (NDiop, n = 50; Dielmo, n = 38; and Diama, n = 46) exposed to distinct densities and proportions of the Anopheles species. Individuals that were not exposed to these tropical mosquitoes were used as controls (Marseille, n = 45).
Methods
Ethics statement
The protocol (N°2006-A00581-50) was approved by the Marseille-2 Ethical Committee (France) and by the Senegal National Ethics Committee (Dakar, Senegal). The written informed consent of each participant was obtained at the beginning of the study, after a thorough explanation of its purpose.
Study sites, sera samples and entomological observations
The study was conducted on two different populations: un-exposed people and people who were regularly exposed to
An. gambiae s.l. and
An. funestus bites
. Forty-five serum samples from French adults living in Marseille (43°17′N, 5°22′E; mean age ± standard deviation (SD): 40.73 ± 12.02, sampled in February 2007[
16]), who had never been in countries endemic for
An. gambiae s.l. and
An. funestus, were used as un-exposed negative controls. The exposed group consisted of 134 individuals living in the Senegalese villages of Diama (16°13′ N, 16°23′W; n = 46, mean age ± SD: 17.96 ± 11.58, sampled between March and October 1994), Dielmo (13°45′N, 16°25′W; n = 38, 28.38 ± 21.26, sampled in March 1995) and Ndiop (13°14′N, 16°23′W; n = 50, 25.87 ± 18.34, sampled between March and December 1995). These populations were exposed to high (Dielmo, approximately a 30.6 human biting rate (HBR), moderate (Ndiop, approximately a 3.9 HBR) and low (Diama, <1 HBR)
Anopheles bite levels. Individuals living in Dielmo and Ndiop were predominantly exposed to
An. funestus (approximately 66%) and
An. gambiae s.l. (approximately 95%), respectively. Individuals living in Diama were predominantly exposed to other anopheline mosquito species (
Anopheles pharoensis, approximately 87%). Details regarding the entomological data are presented in Table
1 and show for each site the amount and the proportion of each anopheline species collected from the three months preceding blood sampling until the end of the blood sampling period. Concerning entomological measures, adult mosquitoes were collected monthly using human bait catches and the HBR, which is the number of mosquito bites per person per night, was calculated as the number of mosquitoes captured during the month divided by the number of person-nights. Additional data on the study site, including entomological and parasitological factors, have been previously reported[
34‐
40].
Table 1
Characteristics of density and exposure to
Anopheles
bites in each site according to entomological data
Diama
|
1,492
| |
<1
| January '94 – December '94 | |
An. gambiae s.l.
| 17 | 1.1% | |
An. gambiae s.s.
| n.d. | n.d. | |
An. arabiensis
| n.d. | n.d. | |
An. funestus
| 0 | 0.0% | |
Others Anophelines (86.9% of An. pharoensis) | 1475 | 98.9% | |
Dielmo
|
1,473
| |
30.6
| January '95 to march '95 | |
An. gambiae s.l.
| 494 | 33.5% | 10.3 |
An. gambiae s.s.
| 13 | 0.9% | 0.3 |
An. arabiensis
| 481 | 32.6% | 10.0 |
An. funestus
| 978 | 66.4% | 20.3 |
Others Anophelines | 1 | 0.0% | 0.0 |
NDiop
|
597
| |
3.9
| January '95 – December '95 | |
An. gambiae s.l.
| 565 | 94.6% | 3.7 |
An. gambiae s.s.
| 190 | 32% | 1.3 |
An. arabiensis
| 373 | 63% | 2.5 |
An. funestus
| 28 | 4.7% | 0.2 |
Others Anophelines | 4 | 0.7% | 0.0 |
Protein expression and purification
The coding sequences for the anopheline proteins gSG6 (gi|13537666), g-5′nucleotidase (gi|4582528), fSG6 (gi|114864550) and f-5′nucleotidase (gi|114864746) were retrieved from the National Center for Biotechnology Information (NCBI) database. The cDNAs of selected proteins were synthesized with a C-terminal His-tag and cloned into the baculovirus expression vector pFast Bac1 (Invitrogen, Cergy Pontoise, France) by Genecust (Gencust, Dudelange, Luxembourg). The fidelity of the cloned sequences was verified by DNA sequencing, using an ABI Prism 3100 analyzer (Applied Biosystems). Recombinant bacmid DNA was generated in the DH10Bac Escherichia coli strain (Invitrogen), using the Bac-to-Bac system and protocol (Invitrogen). Spodoptera frugiperda (Sf9) cells were transfected with the recombinant bacmid DNA using the Lipofectin transfection reagent (Invitrogen), according to the manufacturer’s instructions. The structures of all the inserts were sequenced for authentic cloning (ABI Prism 3100 analyzer, Applied Biosystems). Confirmed clones were amplified in Spodoptera frugiperda (Sf9) cells in serum-free medium (Sf-900 II SFM, Gibco, Carlsbad, CA) to produce working viral stocks, which were titrated by a plaque assay and used for subsequent expression studies. For protein production, Sf9 cells were cultured at 28°C in 800 ml of suspension culture (1,3×106 Sf9 cells/ml), infected with a multiplicity of infection of 5 and harvested after three days by centrifugation at 500 × g for 15 min at 4°C, washing in PBS and repeating centrifugation. All pellets were stored at −80°C until use. The cell pellet was resuspended in lysis buffer (50 mM Tris, pH 7.9, 300 mM NaCl, 20 mM imidazole, with proteolysis inhibitor P8849 (Sigma), and disrupted using an Emulsiflex C3 cell disruptor (Avestin, Mannheim, Germany). The cell lysate was clarified by centrifugation at 40,000 rpm for 45 min at 4°C (45 Tirotor, Beckman Coulter). The recombinant proteins were purified using HisTrap HP columns (AKTA purifier 10GEH, GE Healthcare, France). The fractions containing the His-tagged recombinant proteins were selected by SDS-PAGE and pooled. To eliminate contaminant proteins, pooled fractions of each recombinant protein were then purified by gel filtration (superdex 75 26/60 column, GE Healthcare). The fractions containing recombinant proteins were identified by SDS-PAGE, pooled and dialyzed against a 40 mM Tris–HCl buffer (pH 7.5). The protein concentration was measured using a Lowry DC Protein assay (Bio-Rad, Hercules, CA, USA). The purity of purified proteins was determined by SDS-PAGE, and the identity was confirmed by mass spectrometry (MS).
SDS-PAGE
Five microgram of each purified recombinant protein were reduced in a Tris buffer containing dithiothreitol (1% w/v, Sigma), boiled for 5 min, and loaded per well onto a 15% polyacrylamide gel before to be separated using a Mini PROTEAN II (BioRad, Hercules, CA, USA). After electrophoresis, gels were stained with Coomassie brilliant blue R-250 (ImperialTM Protein Stain, Thermo Scientific, Rockford, IL, USA) and scanned with a high-resolution densitometer scanner (Image Scanner 3, GE Healthcare) and densitometry profiles were analysed using the ImageQuantTM TL software (GE Healthcare). Protein bands from gels were excised for further identification by mass spectrometry. Molecular weights were estimated by comparison with standard molecular weight marker (Biorad).
In-gel digestion and mass spectrometry analysis
Excised bands were digested overnight at 37°C with sequencing-grade trypsin (12.5 μg/mL; Promega Madison, WI, USA) in 50 mM NH4HCO3 (Sigma). The resulting peptides were extracted with 25 mM NH4HCO3 for 15 min, dehydrated with acetonitrile (ACN) (Sigma), incubated with 5% acid formic (Sigma) for 15 min under agitation, then dehydrated with ACN, and finally completely dried using a SpeedVac. The samples were then analysed on a NanoLC-LTQ-OrbitrapVelos-ETD (Thermo Scientific, Bremen, Germany) for identification.
Enzyme-Linked ImmunoSorbent Assays (ELISA)
ELISA was performed according to standard procedures. Microtiter 96-well plates (Nunc Maxisorp Immunoplates, Denmark) were coated for three hours at 37°C with 10 μg/ml (50 μl/well) of either gSG6, g-5′nucleotidase, fSG6 and f-5′nucleotidase purified recombinant anopheline proteins in 0.1M bicarbonate coating buffer (pH 9.6) (Sigma). Three washes were done with 200 μL of PBS (pH 7.4, Sigma, USA) plus 0.05% Tween-20 (Sigma) between each incubation. Plates were blocked overnight at 4°C with 200 μL of blocking buffer consisting of PBS 0.05% Tween and 5% skimmed milk (Beckton, Dickinson Bioscience, USA). Serum diluted 1:50 in blocking buffer was added (50 μl/well) and incubated at 37°C for 1 h. Fifty microliters of horseradish peroxidase (HRP)-conjugated rabbit anti-human IgG (1:10,000, Invitrogen, Rockville, USA) diluted in the blocking buffer were incubated for 1 h at 37°C. Enzyme activity was detected by incubation with 50 μl of tetramethylbenzidine substrate (KPL, USA) for 10 min at room temperature in the dark. The reaction was stopped using 50 μl of 1 M H2SO4. The optical density (OD) at 450 nm was determined with a microplate reader (Versa Max® Turnable Multiplate Reader, Molecular Devices, UK). Each serum was tested in duplicate against the different recombinant antigenic proteins and also without antigen (coating buffer only). In order to improve result consistencies, sera from different study sites have been randomly loaded on each plate and each individual serum were tested on the same plate against the four recombinant proteins. A pool of eight sera from individuals living in Dielmo and Ndiop sites presenting high level of antibody responses against the four anopheline recombinant proteins tested (selected based on ELISA optimisation tests), were used as a positive control on all plates coated. Only plates presenting inter-assay variations in absorbance values of positive controls lower than 20% were included in the analysis. The levels of IgG antibodies were expressed as adjusted OD (aOD), which was calculated for each serum sample duplicate as the mean OD value of antigenic proteins-coated wells minus the mean OD value of the background control wells (i.e., coating buffer without antigenic proteins). Sera whose duplicates showed a coefficient of variation (CV) ≥20% were not included in the analysis. The mean aOD of unexposed individuals plus three standard deviation (SD) was used as cut-off value for seropositivity. Seroprevalence was defined as the pourcentage of seropositive individuals in each group.
Protein sequence analysis
Culicidae protein sequences that were related to the 5′nucleotidases from both
An. funestus (gi|114864746) and
An. gambiae (gi|4582528) were retrieved from the National Center for Biotechnology Information (NCBI) using the BLASTp program. The hit sequences exhibiting a significant alignment (E-value<1.10
-4) and a hit sequence with coverage ≥70% and identity ≥ 50% were selected for further protein sequence comparisons. Multiple sequence alignment was performed with the Clustal W 1.7 multiple sequence alignment program[
41], which is included in the Molecular Evolutionary genetic Analysis 5 (MEGA 5) programs package[
42].
Statistical analysis
After verifying that values in each group did not assume a Gaussian distribution, the Kruskal-Wallis test was used for multiple group comparisons and the Friedman test was used to compare observations repeated on the same subjects in each study site. Two independent groups were compared by the Mann–Whitney U test. The Wilcoxon signed-rank test was used for comparison of two paired groups. Frequencies were compared by the Pearson’s Chi-squared test and Spearman’s rank correlation coefficient was computed when appropriate. All differences were considered significant at p <0.05 and statistical analysis and figures were performed using the computing environment R (R Development Core Team, 2012).
Conclusion
Understanding the complexity of the
Anopheles species behavior is of major importance in vector control interventions to protect human populations against malaria.
Anopheles gambiae s.l. and
An. funestus share particularly anthropophilic tendencies that contribute to their vectorial capacity[
56]. However, the existence of ecological and behavioral differences between these species have important epidemiological consequences[
57]. Indeed, during the dry season, the densities of
An. gambiae s.l. declined in some Sub-Saharan Africa areas, whereas the
An. funestus abundance remained maximal, extending the period of malaria transmission. Therefore, the determination of human exposure to malaria vectors at the quantitative level (
e.g., mosquito bite densities) and at the qualitative level (
e.g., which mosquito species bite humans) should help to adapt malaria control strategies according to the spatial and temporal density of mosquito fauna[
58].
To this end, the analysis of the human antibody response against mosquito salivary antigens has proved to be a relevant tool to assess host/vector contact[
16,
17,
59]. As some areas can exhibit a high biodiversity in terms of mosquito species, a high level of specificity is necessary to assess individual exposure by immunological tests based on mosquito saliva. In addition, the presence of a diverse degree of salivary antigen cross-reactivity between different vector species demonstrates the need to precisely define antigenic candidate biomarkers to reflect exposure to several
Anopheles species and to distinguish vector exposure at the species level[
9].
Therefore, the production of specific mosquito saliva antigens in a recombinant form or by using synthetic peptides is a promising alternative strategy for producing safe and highly standardized antigens on a large scale[
21,
24]. A gain of specificity could be achieved by the use of synthetic peptides that do not share sequence homology with other hematophagous arthropod species[
40]. However, the production of whole recombinant antigenic proteins could be more efficient to detect mosquito exposure[
25].
The well-conservedSG6 protein family within the
Anopheles Cellia subgenus made the SG6 proteins the first anopheline salivary candidates tested for the exploration of the relationship between levels of anti-gSG6 IgG responses and individual exposure to
Anopheles bites. Encouraging data has accumulated for the use of the gSG6 salivary protein as serological marker of exposure to the
Anopheles genus. Incontestably, the gSG6 protein can be used to detect exposure to
Anopheles bites, even in areas of low exposure[
23]. In the present study, the IgG response against SG6 from
An. gambiae and
An. funestus could distinguish individuals un-exposed to
Anopheles bites from individuals exposed to
Anopheles bites and could also distinguish high levels of exposures, independent of the anopheline
Cellia species fauna. However, gSG6 was found to elicit a higher level of response than fSG6 orthologs. Overall, the analysis of IgG responses against SG6 ortholog proteins sustains the use of SG6 proteins as consistent indicators of exposure to three major malaria vectors in tropical Africa (
i.e., An. gamgiae, An. arabiensis and
An. funestus).
In contrast, the presence of significantly different patterns and intensities of IgG responses against 5′nucleotidase anopheline orthologs supports the idea that each of these proteins should possess specific antigenic epitopes. Moreover, the IgG response level against the f-5′nuc protein seems to be associated with An. funestus densities. These initial tests provided encouraging preliminary information on the immunogenicity of the anopheline 5′nuc proteins and present the promising possibility of using the 5′nucleotidase salivary protein from An. funestus as the first species-specific antigenic marker of exposure. Complementary studies are needed to confirm the present assumption.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AL and RC conceived and designed the experiments. ZA and BM performed the experiments. AL, FA, RC, FC and PF analysed the data. BN, AS, BY, RF, VHV contributed reagents/materials/analysis tools. AL, FC, FA and RC wrote the paper. All authors read and approved the final manuscript.