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.

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 35 36 37 38 39 40 .

n.d.: not determined, HBR: human biting rate.

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).

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 (Imperial^TM 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 ImageQuant^TM 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).

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.

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.

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 .

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).

Members of the 5′nucleotidase/Apyrase family have been ubiquitously found in the saliva of hematophagous arthropods 19 . These enzymes are known to facilitate the acquisition of blood meals by removing and degrading pharmacologically active nucleotides (i.e., ATP/ADP) that are important for platelet aggregation at the site of the injury 43 . Three members of the 5′nucleotidase gene family were described at the transcript level in An. gambiae salivary glands 44 . Among them, the putative 5′nucleotidase (gi|4582528) from An. gambiae (g-5′nuc) was highly expressed compared to the other two 5′nucleotidase family members. The identification of the putative 5′nucleotidase at the protein level in An. gambiae saliva (64 kDa) confirmed that this salivary protein could be injected into the host during blood feeding 45 . Moreover, the induction of host antibody responses against members of the 5′nucleotidase/Apyrase protein family has previously been reported 46 47 48 . Collectively, these data suggested that 5′nucleotidase proteins could be used as potential immunological exposure markers to mosquito bites. To determine whether members from this protein family could be selected as antigenic candidates for the evaluation of exposure to mosquito bites at the species level, the degree of protein sequence diversity in the 5′nucleotidase/Apyrase protein family among species from Culicidae was evaluated.

Among 5′nucleotidase anopheline proteins, the 5′nucleotidase from An. funestus (f-5′nuc; gi|114864746) was selected based on its protein sequence peculiarities. Indeed, alignment of this sequence with the An. gambiae putative 5′nucleotidase (g-5′nuc; gi|4582528) indicated that the protein from An. funestus lacks a large part of the amino-terminal domain (Additional file 1). The molecular weights and sizes (a.a. length) of f-5′nuc protein sequences (17 kDa, 139 a.a.) differ from those of g-5′nuc (64 kDa, 570 a.a.), which leads to a low coverage sequence (31%) and a moderate identity (<75%). When blasting the f-5′nuc protein sequence (gi|114864746) against the Culicidae protein database on NCBInr (NIH, Bethesda), 13 protein hits containing homologous sequences could be retrieved (E-value<1.10^-4, coverage ≥70% and identity ≥ 50%). An alignment of the f-5′nuc protein sequence (gi|114864746) with the members of the 5′nucleotidase/Apyrase family from other Culicidae indicated that orthologous protein sequences from the Anopheles and Aedes genera presented amino-acid sequence similarities (Figure 1). Among these homologous proteins, two 5′nucleotidase/Apyrase proteins from An. gambiae (gi|4582528 and gi|58377530) and one from Anopheles stephensi (gi|27372911) presented amino-acid identities between 72-75%, and the identity levels drop to less than 56% for all of the other selected proteins, suggesting an important diversity in this sequence among Culicidae. Interestingly, among all the sequences retrieved, only two members of the 5′nucleotidase/Apyrase family (gi|208657657, gi|208657659) from the Anopheles darlingi species were also truncated at their amino-terminal end, but these proteins presented moderate identities (<55%) with the 5′nucleotidase from An. funestus. Thus, taken together, the f-5′nuc protein peculiarities (i.e., a moderate level of sequence conservation within the Culicidae family, an amino-terminal end truncation, the high abundance of these family proteins in the anopheline saliva and their involvement in eliciting host antibody responses) have highlighted our interest in assessing this protein as a potential marker candidate to distinguish exposure to An. funestus from exposure to An. gambiae s.l..

Concerning g5′nuc (gi|4582528), a sequence alignment showed that members of the 5′nucleotidase/Apyrase family share only a moderate level of conservation of the protein sequences (<66% identical amino acids residues), with the exception of a paralogous protein from An. gambiae (gi|58377530) and a salivary apyrase from An. stephensi (gi|27372911) that possess 99% and 80% identical amino acid residues, respectively (Additional file 2). Due to their determinant role in blood feeding, these enzymes could be submitted to environmental pressures that lead to heterogeneous and independent sequence evolution.

SG6 proteins (gSG6, gi|13537666 and fSG6, gi|114864550) were previously described to be restrictively present in the Cellia subgenus and are well conserved within this anopheline subgenus (Additional file 3) 21 . SG6 recombinant proteins were thus used as a reference immunological tool to evaluate the level of exposure to anopheline mosquitoes.

A poly-histidine tag was added to the C-terminus of each recombinant protein to ensure the purification of whole recombinant proteins. gSG6 and g-5′nuc salivary proteins from An. gambiae, and fSG6 and f-5′nuc salivary proteins from An. funestus were produced in Sf9 cells using a baculovirus expression system and were purified by affinity chromatography (HisTrap HP) and a gel filtration column. The fractions corresponding to recombinant proteins were separated by SDS-PAGE, and a representative purified fraction for each protein is presented in Figure 2. As expected, protein bands with apparent molecular weights corresponding to gSG6 and fSG6 protein fractions were detected at approximately 13 kDa, and fractions corresponding to g-5′nuc and f-5′nuc were detected at approximately 65 kDa and 17 kDa, respectively. Proteins bands of interest were excised from the gel and submitted to mass spectrometry analysis, which confirmed that the detected bands corresponded to the expected recombinant proteins (Table 2). Protein profiles were analyzed using ImageQuant^TM TL software to determine the relative abundance of each purified recombinant protein as previously described 49 50 . The purity of the different recombinant proteins was greater than 90%, and the protein fractions were considered sufficiently pure for ELISA experiments.

^a The band number corresponds to the same Roman numeral numbers as indicated in Figure 2.

The identities of the bands, their NCBi accession numbers, the theoretical molecular weight values, as well as the number of peptide sequences identified, the corresponding percentage sequence coverage and the Mascot score are listed for MS/MS analysis (Individual ions scores > 16 indicate identity or extensive homology and were considered to be significant (p<0.05).

Recently, Rizzo and collaborators evaluated the IgG responses in individuals living in a malaria hyperendemic area of Burkina Faso against recombinant gSG6 and fSG6 proteins from An. gambiae and An. funestus, respectively 26 . These individuals, exposed predominantly to An. gambiae s.l. bites, produced a comparable immune response against both SG6 salivary proteins, indicating a wide cross-reactivity between these anopheline orthologous proteins. The high conservation of the gSG6 sequence between members of the An. gambiae species complex (greater than 99%) and members of the Cellia subgenus, such as An. funestus (with an identity of approximately 80%), supports this idea of shared-antigens 21 25 26 . Based on these results, the gSG6 salivary protein was considered to be a valuable serological marker of exposure to Afrotropical malaria vectors.

However, Rizzo and collaborators did not evaluate the IgG responses against these two orthologous SG6 proteins in individuals exposed predominantly to An. funestus mosquito bites. To determine whether variations in the anopheline populations and densities could influence the IgG responses against recombinant orthologs of SG6 proteins from An. gambiae and An. funestus, the IgG responses against recombinant gSG6 and fSG6 proteins were assessed in our study in 134 individuals living in three Senegalese villages, Diama (n = 46), Ndiop (n = 50) and Dielmo (n = 38) which are considered to be weakly (human biting rate (HBR) < 1) 37 38 39 , moderately (HBR = 3.9) 35 36 and highly (HBR = 30.6) exposed to Anopheles bites 34 , respectively. These individuals were exposed to distinct anopheline populations, with a predominance of An. pharoensis (approximately 87%) and An. gambiae s.l. mosquitoes (greater than 94%) in Diama and NDiop, respectively, and a majority of An. funestus mosquitoes (66.4%) in Dielmo (Table 1). Sera from European individuals (n = 45) living in Marseille, collected in February 2007 and never exposed to these tropical malaria vectors, were used as un-exposed controls. The IgG response against the gSG6 and fSG6 proteins was assessed by ELISA.

For gSG6 protein, the IgG responses were significantly different among the four groups (Kruskal-Wallis test, p<0.0001, Figure 3A). When comparisons were performed between two sites, the anti-gSG6 IgG responses were significantly higher in exposed individuals from each village (i.e., the means of aODs with a 95% confident interval (95%CI) were +0.50 [0.39 to 0.60], +0.65 [0.53 to 0.76] and +0.58 [0.49 to 0.66] for Diama, Dielmo and NDiop, respectively) compared with un-exposed control individuals (+0.16 [0.12 to 0.20]; Mann–Whitney U test, p<0.0001). Conversely, despite distinct exposure levels to Anopheles bites in these three Senegalese villages, significant differences were detected only between Diama and Dielmo (Mann–Whitney U test, p=0.0174). Although the highest exposure to Anopheles bites was in Dielmo, followed by NDiop and finally Diama; the seroprevalence (a seropositivity cut-off was set as the mean aOD of un-exposed controls + 3 SDs: 0.164 + (3 x 0.130) = 0.554), with values of 55%, 52% and 37% in Dielmo, NDiop and Diama, respectively, was not found to be significantly different among the three villages (Pearson’s Chi-squared test, ns, Figure 3C). These data support the hypothesis that gSG6 appears to be a valuable marker for distinguishing individuals who are un-exposed to anopheline bites from individuals who have been exposed to anopheline bites, and that this candidate is also well-adapted for the detection low levels of exposure to Anopheles bites 20 23 . Although the anti-gSG6 IgG response could discriminate exposed individuals from “low” to “high” Anopheles bite densities, such as in Diama and Dielmo, it seems to be less capable of distinguishing “moderate” exposure, such as in NDiop, compared with high and low exposure sites. The individual heterogeneity of the IgG responses against gSG6 in each site could explain this powerless discrimination. This phenomenon could be attributed to the heterogeneity of Anopheles exposures according to distinct human behaviors or attractive differences, as recently described 51 52 . To better appreciate the variations of anti-gSG6 IgG levels, according to Anopheles densities, a kinetic analysis of the IgG response may be more appropriate and necessary.

For the fSG6 protein, significantly different IgG responses were found among the four groups (Kruskal-Wallis test, p<0.0001, Figure 3B). These significant differences were attributed to higher IgG responses from exposed individuals living in Dielmo (i.e., a mean aOD [95%CI] of + 0.38 [0.29 to 0.48], Mann–Whitney U test, p<0.0001), NDiop (i.e., + 0.35 [0.26 to 0.44], Mann–Whitney U test, p<0.0004) and Diama (i.e., + 0.27 [0.19 to 0.34], Mann–Whitney U test, p=0.0337) compared with un-exposed control individuals (i.e., + 0.15 [0.10 to 0.19]). As observed for gSG6, a comparison of the IgG responses between two Senegalese villages, indicated a significant difference only between Diama and Dielmo (Mann–Whitney U test, p=0.0194). The seroprevalence from Diama (17%), Dielmo (11%) and NDiop (22%) (the seropositivity cut-off was set as the mean aOD of un-exposed controls + 3 SDs: 0.148 + (3 × 0.146) = 0.586) was not found to be significantly different among the three villages (Pearson’s Chi-squared test, p>0.05, Figure 3D). As observed for gSG6, the anti-fSG6 IgG responses increased significantly in populations exposed to Anopheles bites, even in areas of low exposure; however, this method seems to be insufficiently sensitive to discern different gradients of Anopheles densities, and the immune response appears not to be perturbed by the anopheline fauna variations.

To estimate the cross-reactivity level of the IgG responses from exposed individuals of the three Senegalese villages (n = 134) against SG6 orthologs, a Spearman’s rank correlation coefficient (rho) test was used, and the corresponding p-values were determined (Figure 4). A significant positive coefficient correlation (r = 0.6208, p < 0.0001) was observed between the IgG responses against g-SG6 and f-SG6 among the exposed individuals. Despite the variations of Anopheles species proportion among the three villages, the large majority of individuals were “equivalently” distributed around the best-fit line. Thus, these observations strongly support a wide cross-reactivity to the gSG6 and fSG6 salivary antigens, which is in agreement with a previous study 26 . In agreement with Rizzo et al., the slope (0.5923±0.0527) was greater than 0.45, and the best-fit line runs below the diagonal 26 . This phenomenon resulted of in a higher IgG response level against gSG6 compared with fSG6 in the three villages, independent of the Anopheles population proportion. Interestingly, the IgG response against the fSG6 protein was significantly lower than that against gSG6, even in areas where individuals are predominantly exposed to An. funestus (Wilcoxon signed-rank test, p<0.0001, Additional file 4). These results suggest that, despite high sequence conservation among these orthologous SG6 proteins, fSG6 appears to be less antigenic than gSG6. The presence of few species-specific epitopes in each SG6 protein could explain these differences in IgG responses 26 .

Collectively, these data uphold that SG6 proteins could be specific markers of exposure to the Anopheles Cellia-subgenus, and the use of these two proteins could improve the determination of the individual exposure level to these malaria vectors 26 .

The presence of numerous Anopheles species exhibiting differences in their biology and behavior and variations in mosquito species proportion and density throughout the seasons could have major implications for vector control programs 53 54 . Moreover, the geo-repartition and the density of the Anopheles species could vary from area to area in relatively close sites 55 , which can increase the complexity of defining the precise regional malaria vectors responsible for malaria transmission. Thus, the achievement of species-specific individual exposure markers from major Afrotropical malaria vectors appears essential to evaluating the appropriateness of vector control measures and to identifying the Anopheles species considered to be the main malaria vector in any area at any given time point.

In addition to the SG6 Anopheline proteins, this study aimed to evaluate f-5′nuc as a possible An. funestus species-specific mosquito-bite-exposure marker. To assess the use of anti-f-5′nuc IgG responses as a tool to distinguish An. funestus exposure from An. gambiae s.l. exposure, the immune responses from individuals of the same Senegalese villages were evaluated by an ELISA against 5′nucleotidase ortholog proteins from these two mosquito species.

For the g-5′nuc protein, the IgG responses were significantly different among the four groups (Kruskal-Wallis test, p<0.0001, Figure 5A). When comparisons were performed between two sites, the anti-g-5′nuc IgG responses were significantly higher in exposed individuals from each village (i.e., the mean aODs [95%CI] were +0.50 [0.40 to 0.61], +0.51 [0.39 to 0.63] and +0.65 [0.52 to 0.77] for Diama, Dielmo and NDiop, respectively) compared with un-exposed control individuals (i.e., +0.16 [0.12 to 0.21]; Mann–Whitney U test, p<0.0001). Conversely, no significant differences in the anti-g-5′nuc IgG responses was found, regardless of which two groups from these three Senegalese villages were being compared (Mann–Whitney U test, ns). Although a higher seroprevalence (the seropositivity cut-off was set as the mean aOD of un-exposed controls + 3 SDs: 0.163 + (3 x 0.147) = 0.604) was detected among individuals living in the NDiop area (40%), where An. gambiae s.l. is largely predominant, compared with Diama (28%) and Dielmo (29%), no significant differences were found among the three villages (Pearson’s Chi-squared test, ns, Figure 5C). These data suggest that the anti-g-5′nuc IgG responses could distinguish individuals un-exposed to Anopheles from individuals exposed to Anopheles, independent of the level of exposure or the anopheline fauna.

For the f-5′nuc protein, the IgG responses were significantly different among the four groups (Kruskal-Wallis test, p<0.0001, Figure 5B). In particular, the anti-f-5′nuc IgG responses were significantly higher for exposed individuals living in Dielmo (i.e., + 0.80 [0.65 to 0.96]) and NDiop (i.e., + 0.63 [0.50 to 0.76]), compared with un-exposed control individuals (Mann–Whitney U test, p<0.0001 for these two comparisons) or with individuals living in Diama (Mann–Whitney U test, p<0.0001 compared with Dielmo and p<0.002 compared with NDiop). No significant difference was noted between individuals living in Diama (i.e., a mean aOD [95%CI] of + 0.34 [0.22 to 0.46]) and un-exposed control individuals (i.e., + 0.22 [0.16 to 0.28]; Mann–Whitney U test, ns). The seroprevalence (the seropositivity cut-off was set as the mean aOD of un-exposed controls + 3 SDs: 0.219 + (3 x 0.210) = 0.849) was found to be significantly different among individuals living in Dielmo (47%) and those living in Diama (17%) (Pearson’s Chi-squared test, p<0.007, Figure 5D), corresponding to areas where An. funestus is predominantly present and absent, respectively. The seroprevalence of individuals living in NDiop (26%) was not significantly different when compared with the two other Senegalese villages (Pearson’s Chi-squared test, ns).

The absence of a significant anti-f-5′nuc IgG response among un-exposed individuals and among individuals living in Diama that were not exposed to An. funestus bites, in addition to the detection of highly significant differences in the anti-f-5′nuc IgG responses between individuals not exposed to An. funestus bites and individuals living in Dielmo or NDiop, suggests that exposure to An. funestus bites seems to be required to elicit an antibody response against the f-5′nuc protein. Indeed, the higher IgG response against f-5′nuc that was detected at a site where An. funestus is predominant is a supplementary argument supporting the idea that this salivary protein could reflect a species-specific exposure. Nevertheless, the data could not exclude the possibility that the anti-f-5′nuc IgG responses could correspond to the level of Anopheles bites. Indeed, the anopheline specimens collected in these three villages suggested a decreasing mosquito density gradient from Dielmo to Diama, with an intermediate exposure level in NDiop, which corresponds to the anti-f-5′nuc IgG pattern observed among these three areas. Comparisons of the anti-f-5′nuc IgG responses from individuals exposed uniquely to An. funestus or to An. gambiae s.l. at equivalent densities could dispel this ambiguity. Nevertheless, paired comparisons of the IgG responses against 5′nuc orthologous proteins indicated a significant decrease and increase for individuals living in Diama (Wilcoxon signed-rank test, p=0.023) and Dielmo (Wilcoxon signed-rank test, p=0.004), respectively (Additional file 4).These data suggest that the recognition of these salivary proteins by the immune system was not equivalent and that specific antigenic epitopes may be associated with each protein.

To assess the cross-reactivity level of the IgG response of exposed individuals from the three Senegalese villages (n = 134) against 5′nuc orthologs, a Spearman’s rank correlation coefficient (rho) test was used, and the corresponding p-values were determined. In contrast to the SG6 orthologs, no significant correlation was observed between the IgG responses from exposed individuals against g-5′nuc and f-5′nuc recombinant proteins (r = 0.1635, p=0.0591). A high dispersion of the dots was visible on the scatter plot (Figure 6), underlining the fact that some individuals presented a larger response to either g-5′nuc or to f-5′nuc proteins. Thus, the antibody response against these 5′nuc ortholog proteins indicated a low cross-reactivity, suggesting that, despite a partial conservation of the carboxy-terminal end of these proteins (74% identical), the two proteins may be harboring “species”-specific epitopes. In addition, the large amino-terminal segment of g-5′nuc (i.e., from amino acid 1 to 434) could possibly possess antigenic sequences that are absent in f-5′nuc and could explain the differences in serological recognition. Moreover, the g-5′nuc amino-terminal sequence possesses several cysteines (n = 7) that may be involved in protein folding, which could change the epitope conformation or mask some epitopes in the carboxy-terminal (which has partial homology with f-5′nuc protein sequence). Indeed, the resulting cryptic epitopes in the g-5′nuc protein could be, in contrast, accessible to an antibody response in the f-5′nuc protein and, therefore, induce a “species”-specific serological response.